Skip to main content

Full text of "General Virology"

See other formats


CO > CO ? 

LIJ< OU 168323 > 

> LT 7) A* 

^99 < co 

^ _l > 


Osmania University Librarq 

Call No. 5&1 W W lC \ 4 ^ Accession No. A- <l 2- { 3 


This book Wld ba returned on or before the date last marked below 

General Virology 

Frontispiece. A hair cell of tobacco, containing a crystal of tobacco mosaic virus, 
is photographed (a) before and (b) after freeze-drying. The crystal is extracted 
by micromanipulation in the frozen-dry state (c). Upon partial dissolution in 
water such crystals release characteristic virus particles (d). From: Steere and 
Williams (625a). Courtesy Dr. R. C. Williams, Virus Laboratory, University of 
California, Berkeley. 

General Virology 

S. Iv 1,1'RIA 

IMtOKI SSOH 01- H\< 'ILHIOI ()(.Y 

I \1\ 1 Hsm OF II LINOIS 






All Rights Reserved 

This book or any part thereof must not 
be reproduced in any ff rn without 
the written permission of the imltlinhw . 


Library of Congress Catalog Card Number: 53-11874 




This book is the outgrowth of a course on viruses that I have been 
teaching since 1946, first at Indiana University, more recently at the 
University of Illinois. The problem that faced me in 1946 was plan- 
ning a course in virology for graduate and advanced undergraduate 
students in biology and biochemistry, who had no medical orientation 
and no background in histopathology, in a university that was justly 
proud of its reputation as a center of experimental biology. I could 
teach either a watered-down course in virus diseases or organize a 
new type of course, in which virology would be presented as a bio- 
logical science, like botany, zoology, or general bacteriology. My 
choice of the second alternative was, I think, justified. My classes in 
virology have been well attended and have attracted excellent students. 
Similar courses have since been established in other institutions. 

Virology is fast becoming an important field of science, in which 
geneticist, cell physiologist, and biochemist find, in the ground plowed 
by the pathologist, a fertile soil for new approaches to fundamental 
problems of cell function and organization. The interest of "outsiders" 
in viruses has grown continuously since the middle 1930*5, stimulated 
at first by the progress of physicochemical work on virus particles, later 
and more powerfully by the recognition of viruses as keys to the study 
of cellular integration. The eagerness with which modern bacteri- 
ophage work has been seized upon by biophysicists, geneticists, and 
biochemists, and the conscious efforts to create a comparative virology 
(whose heuristic value can be seen by perusing, for example, the pro- 
ceedings of a symposium on The Nature of Virus Multiplication held 
at Oxford in 1952) are signs of the healthy growth of the new science. 

In attempting to teach general virology, I was faced with the prob- 
lem of the lack of a textbook. In 1946, the best books dealing with 
viruses were devoted to the description of virus diseases. The main 
exceptions were Doerr and Hallauer's Handbuch der Virusforschung, 
a bilingual treatise of vast scope, and Bawden's excellent Plant Viruses 
and Virus Diseases. Both were unsuitable as textbooks, although in- 
valuable as reference books. In spite of many important additions to 
virological literature, no single volume suitable for classroom use has 


x Preface 

appeared. This book is an attempt to fulfill the need for a textbook 
in general virology. 

My teaching of virology, and this book as a result of it, have been 
built around a central concept, that of the dual nature of viruses as 
inert particles on the one hand, and as operating constituents of func- 
tional cells on the other hand. In the light of this concept, I have 
tried to present the physical and chemical approach to virus particles 
and the biochemical and cell-physiological approach to virus-infected 
cells as two separate but integrated aspects of virology. In the limited 
space of one semester or of a book of this size, one can hardly hope 
to make a biochemist or a physicochemist out of a biology student, or 
a biologist out of a chemist. I have covered such background infor- 
mation as I found useful in teaching, from the logarithms to the em- 
bryology of the chick embryo, but I have given no details of actual 
techniques. Any description of individual virus diseases has deliber- 
ately been omitted. The pathogenicity of a virus is, of course, less 
incidental to virus biology than, say, the pathogenicity of Strepto- 
coccus pyogenes is to bacteriology, since the virus, an integrated intra- 
cellular parasite, "lives" the life of the host as its own only life. Thus, 
each virus disease is potentially a different form of virus life. Yet, I 
feel justified, both on didactic and on conceptual grounds, in assuming 
a priori as much uniformity and community of mechanisms as the 
known facts do not contradict. This assumption stems, of course, 
from a belief in the intrinsic simplicity of nature and from a feeling 
that the ultimate contribution of science resides in the discovery of 
unifying and simplifying generalizations, rather than in the description 
of isolated situations in the visualization of simple, overall patterns, 
rather than in the analysis of patchworks. 

Facts about individual viruses are presented as model systems or as 
examples, without any attempt at extensive coverage. The choice of 
examples reflects my limited knowledge of many areas of virology. 
Being myself a specialist, I shall count on the tolerance of other special- 
ists whose field I may have misinterpreted, and on their willingness to 
suggest improvements for future editions of this book. The selection 
of references was dictated not only by the accidents of my limited 
knowledge but also by an attempt to single out articles with further 
references, with descriptions of important methods, or with new ideas 
and timely syntheses. 

In a science developing as fast as virology, any book is bound to 
be somewhat out of date by the time it appears in print. In fact, 
original work is proceeding at such a pace that interpretations pre- 

Preface xi 

seated in research articles must often be revised in galley proof. Yet, 
virology has reached the stage where we may be justified in attempting 
at least a provisional integration. It will be gratifying if this book 
contributes to such an integration. 

Many friends and colleagues encouraged me to write this book; 
friendly periodic reminders from the publishers over a period of three 
years helped me fight the frequent temptation to forget about it. 
Special thanks go to Dr. Zella Luna, who read the whole manuscript 
and contributed many improvements of language, style, and reasoning, 
and to Mrs. Mary Delbriick, who in the summer Of 1949 typed under 
dictation the first draft of several chapters. My friends Drs. L. M. 
Black, E. Caspari, M. Delbruck, G. K. Hirst, K. Maramorosch, 
S. M. Rose, R. W. Schlesinger, and R. Y. Stanier read some chapters 
at various stages of writing and made valuable suggestions. The 
Graduate School of Indiana University provided in 1949 a grant for 
secretarial help. I wish especially to thank my very good friends the 
students in the virology courses at Indiana University and at the Uni- 
versity of Illinois, who stimulated and shared my enjoyment of virology 
and submitted themselves sympathetically to my early and recent ex- 
periments in developing an approach to this growing science. 

S. E. L. 

November, 1953 


1 Introduction The Science of Virology 1 

2 Detection and Identification of Viruses 20 

3 Titration of Viruses 39 

4 Size and Morphology of Virus Particles 55 

5 * Purification and Chemical Composition of Virus Material 85 

6 Serological Properties of Viruses 116 

7 Environmental Effects on Virus Particles 135 

8 Virus-Host Interaction The Bacteriophage-Baeterium 

System 157 

9 The Bacteriophage-Baeterium System (Continued) 183 

10 The Interaction of Plant Viruses with Their Host Plants 209 

11 The Interaction of Animal Viruses with Their Hosts Tissue 

Cultures Intracellular Inclusions 220 

12 Growth of Viruses in the Chick Embryo 239 

13 Hemagglutination Phenomena and Virus Growth Sum- 

mary on Virus Reproduction 250 

14 Interference Phenomena in Virus Infections 273 

15 Variation in Viruses Host Variation and Susceptibility to 

Viruses 290 

16 Transmission, Vectors, and Survival of Viruses 307 

17 Viruses and Tumors 321 

18 Origin and Nature of Viruses 344 

19 Appendix-The Rickettsiae 364 

Bibliography-Author Index 375 

Index 415 


"For understanding life phenomena it is 'neither sufficient to know 
the individual elements and processes nor to interpret their order by 
means of machine-like structures, even less to invoke an entelechy as 
the organizing factor. It is not only necessary to carry out analysis 
in order to know as much as possible about the individual components, 
but it is equally necessary to know the laws of organization that unite 
these parts and partial processes and are just the characteristic of vital 
phenomena. Herein lies the essential and original object of biology." 

Problems of Life 


The Science of Virology 


Virology has become a fundamental biological science in its own 
right. Just as bacteriology has emerged as a biological science out of 
the practically important but scientifically constricting borders of its 
medical applications, so has virology begun to become a body of 
knowledge and of generalizations, with its own perspectives and its 
own internal development. Having originated as a branch of pathol- 
ogyhuman and animal pathology on the one hand, plant pathology 
on the other hand the new science of virology, developed at first in 
response to practical needs, has reached a point where progress is 
dictated at least as much by the logic of its internal development as 
by the demands of applied areas. Analogy with other fields teaches 
us that the emergence of virology as a fundamental science from an 
applied one will actually make virology more adequate to handle 
those practical problems from which it arose, even though it may some- 
times appear to lose sight of them. The years between 1945 and 1950 
saw an increasing integration of various areas of virology, particu- 
larly under the impact of advances in the study of bacterial viruses 
(172). Nevertheless, the attempts to present virology in a coordinated 
way, or at least in an all-inclusive way, have been few (187; 610). 
Still today, the methodological and semantic barriers between plant, 
animal, and bacterial virologists are slow in yielding before the recog- 
nized need for joint efforts and for cross-fertilization of ideas. 

What kind of biological science is virology? We may subdivide 
somewhat artificially the fundamental biological sciences (as distinct 
from applied sciences, like medicine, and from ancillary sciences, like 
biometrics) according to the nature of their subject matter, into taxo- 
nomic, integrative, and interpretative sciences. A taxonomic science 


2 The Science of Virology 

(for example, botany, mycology, entomology, ichthyology, mammal- 
ogy) is characterized- by the fact that its subject matter is a group of 
organisms with a recognized taxonomic unity, that is, a common an- 
cestry and a historical development unique to that group. Integrative 
sciences (physiology, ecology, genetics) analyze the common or 
specialized properties of living organisms in their historical dynamic 
relations and transformations. Interpretative sciences (biochemistry, 
biophysics) analyze elementary processes and functions of organisms 
in terms of the behavior of the pieces (molecules, atoms, electrons) 
that are the common material basis of all matter, living and nonliving. 
A definition of viruses. Does virology fit into any of the above 
categories? This question has no precise answer. The subject of 
virology is not one immediately definable by common-sense criteria 
verified by taxonomic or methodological analysis. Viruses, the subject 
matter of virology, themselves require a definition. This, like all defi- 
nitions, should be operational, that is, it should provide factual criteria 
for inclusion or exclusion of given objects in terms of observable prop- 
erties and performable tests. Such a definition always has a certain 
arbitrary quality. Its value will depend on the number and size of 
the areas of uncertainty it leaves. 

We shall adopt for viruses the following definition: Viruses arc sub- 
microscopic entities, capable of being introduced into specific living 
cells and of reproducing inside such cells only. This definition provides 
practical, restrictive criteria, and at the same time emphasizes the fact 
that virology, although it covers a group of biological entities, is not a 
taxonomic science in the usual sense. Indeed, there are no grounds 
for assuming that all objects that fulfill our definition belong to one 
distinctive branch of the evolutionary tree. There is not even general 
agreement as to whether any or all viruses can be considered as 
"organisms." Our definition of viruses stresses the methodological 
rather than taxonomic unity of the subject matter of virology. Yet it 
stresses the fact that viruses, although possibly not constituting a taxo- 
nomically valid group, possess the basic properties that can be validly 
accepted as an operational basis for defining "organisms" the proper- 
ties of individuality and of homologous reproduction. Reproduction- 
together with its inseparable counterpart, heritable variability makes 
possible the historical continuity and perfectibility of the pattern of 
specificity embodied in the individual and gives it potential immortality. 

We must point out, however, that the reproduction referred to in 
our definition is purely descriptive, stating that more virus similar to 

CH. 1 Virology as a Biological Science 3 

the original one is produced. No specific mechanism of reproduction 
is postulated in the definition. 

Let us examine our definition more closely. We shall discuss the 
meaning of its various parts and see how the elimination of any one 
of them suggests possible natural relationships between viruses and 
other biological elements. 

The requirement for ability to be introduced into host cells em- 
phasizes the external derivation of the virus. This constitutes a re- 
quirement for the recognition of a virus as such. A virus does not 
need to enter every host from outside. It may be transmitted internally 
from generation to generation of its host, even intracellularly at cell 
division and in the formation of germ cells. But, to be observed, a 
virus must be capable at some time of entering some host organism or 
cell from the outside. Elimination of this requirement would identify 
viruses with all "self-reproducing" protoplasmic components of cells, 
such as genes and other units endowed with genetic continuity. As 
we shall see in chapter 18, the view is rather widely held that some or 
even most viruses may have originated by the acquisition of infectivity 
(that is, of external transmissibility ) on the part of self-reproducing 
cell components. At any rate, we may emphasize from the start that, 
once inside the host cell, a virus appears indeed to behave as a proto- 
plasmic element, distinct, however, from other such elements by its 
actual or potential transmissibility to new host cells. 

The "submieroscopic" requirement is more arbitrary, but is method- 
ologically convenient. The effective resolving power of the light 
microscope being around 2000 A, the definition restricts the virus field 
to the study of agents that at some time in their development consist 
of elements, recognizable in isolated form, with at least one linear 
dimension equal or smaller than 2000 A (48). There is, of course, 
no fundamental reason behind the choice of this borderline value for 
size. It simply turns out empirically to be an adequate point of sepa- 
ration. This size limit happens to be reasonably close to the limit of 
porosity of ordinary bacteriological filters, which can therefore be used 
to separate bacterial cells from virus particles without much loss of 
the latter (filtrable viruses). Several agents, however, which have 
dimensions greater than 2000 A are included among viruses. Our 
definition should really state: "submicroscopic or nearly submicro- 
scopic entities." 

Elimination of the submicroscopic size requirement would include 
among viruses a variety of obligate intracellular parasites, such as some 
bacteria (e.g., Mycobacterium leprae), the rickettsiae, and some algae 

4 The Science of Virology 

and possibly fungi. There are sound taxonomic reasons for including 
such organisms with groups other than viruses. The rickettsiae (see 
chapter 19), according to the most accepted view, represent a special 
group of obligate parasitic Gram-negative bacteria. The possibility 
that some of the obligate parasitic microorganisms are related to viruses 
and the hypothesis that some viruses originated from them or from 
their ancestors by "regressive evolution" through parasitism are very 
popular among virologists (114; 271; 395). 

The requirement for reproduction "inside living cells only'* excludes 
all saprophytic, free-living organisms. A number of submicroscopic 
free-living organisms are known to exist. Many bacteria, especially 
in unfavorable environments, can go through submicroscopic stages 
or L forms, which may reproduce as such and later return to the 
typical bacterial morphology (179). Several submicroscopic free- 
living organisms, without known bacillary stages of development, have 
been described. Among them are the "pleuropnetimonia organisms" 
originally described by Nocard and Roux ( see 571 ) , similar in many 
respects to L forms of bacteria; the sewage and soil organisms of 
Laidlaw and Elford (396) and of Seiffert (589); and the serum 
organisms of Barnard (39). The pleuropneumonia and sewage or- 
ganisms have well-defined and not too complex nutritional require- 
ments for growth (519). This in itself differentiates them from viruses, 
whose growth, as we shall see, appears to depend on the host cells not 
for a supply of nutritionally required compounds or growth factors, 
but for the use of the integrated enzymatic machinery of the cell, which 
provides energy and synthetic machinery for the virus. Indeed, viruses 
in the free state appear to be completely inert metabolically. 

The requirement for "host specificity" included in our definition of 
viruses, although not excluding any known group of organisms, em- 
phasizes again the fact that the virus-host relation is one of integration 
rather than of supply of nutrients. If we were to encounter an intra- 
cellular parasite apparently unable to reproduce in the free state but 
capable of reproduction in living cells of any kind, we could reasonably 
suspect that growth of this parasite in the free state would be possible 
if we were able to isolate and supply in cell-free form some hitherto 
unidentified, perhaps unstable, nutrient. On the other hand, we shall 
see that the relation of true viruses to their host cell is so intimate and 
integrative that the hope for cell-free virus reproduction is about on 
the same level as the hope for artificially constructed, self-reproducing 

CH. 1 Development of Our Knowledge of Viruses 5 

The relation of virology to other biological sciences. Virology's 
relation to bacteriology stems on the one hand from the common 
technical problems of handling very small biological objects (micros- 
copy, filtration techniques, sterilization), and on the other hand from 
the common interest in pathogenic microbes. Both pathogenic bac- 
teriology and applied virology belong to the wider field of pathology. 
The study of pathological changes in host organisms, however, is more 
intimately connected with the study of viruses than with that of patho- 
genic bacteria, because the detection, the recognition, and the titration 
of viruses depend almost exclusively on observations of abnormal 
changes produced in some host. Fundamentally, however, virology 
should be concerned primarily with virus properties and functions. It 
should ultimately be possible to interpret all pathological changes of 
a host, directly or indirectly, in terms of the mechanisms by which a 
virus alters the infected cells. 

Virology has become closely allied to protein chemistry and physico- 
chemistry and has borrowed the techniques of these sciences, because 
the small si/e of viruses places them in the colloidal range and gives 
them many properties in common with proteins and other macro- 
molecular substances. Methods for purification and for determination 
of the size, homogeneity, and state of dispersion of particles are similar 
for viruses and proteins. The overlapping size ranges of viruses and 
proteins do not a priori imply a similarity of organization or of chemical 
complexity. Such a similarity can only be tested by structural analysis. 
Thus, the relation of virology to protein chemistry is, at least in prin- 
ciple, purely technological. 


Like all sciences, virology has not developed in a straight path, but 
rather by a slow accumulation of empirical knowledge. Some unity 
and general perspectives have only emerged in the last 10 years. Some 
diseases now known to be caused by viruses have been recognized for 
thousands of years. A Chinese description of a pestilence dating from 
the 10th century B.C. apparently refers to smallpox. Yellow fever, 
known for centuries in tropical Africa and as a scourge of ships in the 
African trade, was probably responsible for the legends of cursed ships, 
such as those of the Ancient Mariner and the Flying Dutchman (114). 
Plant virus diseases, such as potato leaf roll, have been traced to records 
of several hundred years ago; and tulips with the ornamentally appre- 

6 The Science of Virology 

dated color variegation known as tulip break, caused by a virus, have 
been cultivated since the 16th century (43). 

The transmissibility of smallpox has been known for centuries, and 
vaccination against smallpox by extracts containing vaccinia virus 
(cowpox) was introduced as a medical practice by Jenner at the end 
of the 18th century ( Jennerian vaccination, 361 ). The transmissibility 
of tobacco mosaic by mechanical inoculation of sap of infected plants 
was demonstrated by Mayer in 1886. 

During the last decades of the 19th century, the successes that had 
attended the search for bacterial agents of many diseases drew in- 
creasing attention to various diseases for which this search had re- 
mained fruitless. The idea of submicroscopic, nonbacterial patho- 
gens, however, was slow in finding an experimental basis. In 1892 
Iwanowsky reported the transmission of tobacco mosaic by means of 
sap filtered through bacteriological filters which were supposed to 
retain all bacteria. His report went unnoticed; its significance was 
apparently not fully clear to the author himself. In 1898-1899 Loeffler 
and Frosch (428) for foot-and-mouth disease and Beijerinck (58) for 
tobacco mosaic succeeded in proving serial transmission by bacteria- 
free filtrates in which no microscopic organism could be detected. 
Impressed by this unexpected finding, Beijerinck described the agent 
of tobacco mosaic as a contagium vivum fluidum, meaning by this an 
agent which reproduced, and therefore had life, but which was in a 
state of dispersion different from that of organisms. In reality, there 
is no clear-cut difference in the state of dispersion of small organisms 
and of large molecules. Moreover, the fact that virus reproduction 
and its pathological consequences can be initiated by a single virus 
particle make the state of dispersion of viruses irrelevant for their 
mode of action. 

There followed an intense period of discovery of virus pathogens or 
"filtrable viruses," to employ a now obsolete expression. The early 
years of virology saw the development of methods permitting the 
microscopic visualization of the largest types of virus particles or ele- 
mentary bodies. Following pioneer observations by Buist (106), many 
workers, and especially Paschen (514), greatly developed the art of 
virus staining. Meanwhile, rapid progress was made in the study of 
the cellular pathology of virus diseases, with the recognition of specific 
intracellular inclusions (see 217). 

On the one hand, the work on the size and properties of virus 
particles contributed to the development of modern techniques of 
ultrafiltration, ultracentrifugation, and other physicochemical proce- 

CH. 1 Plan of the Book 7 

dures, culminating in the successful purification, crystallization, and 
chemical characterization of virus particles (405; 622). On the other 
hand, research on cellular pathology of virus infections gave a great 
stimulus, first, to the study of tissue cultures, and second, to the study 
of chick-embryo techniques as means of investigating viruses in sim- 
plified systems. vThe study of tissue cultures provided direct evidence 
of the need for contact with living, metabolizing cells as a prerequisite 
for virus reproduction. Frozen, killed tissue, or tissue separated from 
virus by membranes impermeable to the virus would not support virus 
reproduction. The discovery in 1910 of a virus that produces malig- 
nant tumors in chickens ( 562 ) and the generalization of this discovery 
to a whole group of fowl cancers in subsequent years opened the way 
for the realization that viruses are a major agent of neoplastic trans- 
formations, both in animals and in plants (83). 

The discovery of bacteriophages or bacterial viruses by Twort (653) 
and by d'Herelle (306) and the deliberate use of bacteriophages as 
models for the study of the virus-cell relation (65) provided perhaps 
the most important single factor for the present integration of virology 
into a unified science. 

The recognition of the specificity of host-virus relations, of its limita- 
tions, and of its determination by genetic and developmental factors 
( host specificity and tissue specificity ) assumed increasing importance 
as a result of efforts to conquer the virus diseases of man and of eco- 
nomically important animals and plants (see 265). From the epi- 
demiological standpoint, the most salient developments have been the 
recognition of the role of arthropod vectors in the transmission of many 
virus diseases of animals and plants (547), the analysis of complex 
host- vector-virus cycles (601), and the clarification of the role of 
latent infections of reservoir hosts in perpetuating pathogenic viruses. 
The analysis of spontaneous mutability in viruses. ( 310; 361 a; 457) not 
only contributed to the epidemiological understanding of virus diseases 
but also established the nature of viruses as independently evolving 
and therefore taxonomically independent genetic systems. Similarly, 
the analysis of the serological properties of viruses (121) established 
their nature as chemically specific, host-independent structural ele- 
ments, while providing a basis for a multitude of diagnostic tests for 
virus diseases. 


It will be our aim to study viruses as a group, in spite of the recog- 
nized uncertainty as to the extent of taxonomic kinship among what 

8 The Science of Virology 

we call viruses. We shall not attempt the description of individual 
virus diseases nor of individual viruses, but shall rather deal with the 
facts and methods of virology as a whole. We shall, however, under- 
take whenever possible the interpretation of certain phases of applied 
virology in terms of fundamental virus properties. For example, a 
discussion of virus ecology will of necessity be closely allied to a 
general discussion of the epidemiology of virus diseases. We shall 
subdivide our subject matter as follows: 

1. Survey of viruses; range of existence; nomenclature and classifi- 
cation (chapter 1). 

2. Detection and titration of viruses (chapters 2 and 3). 

3. The properties of viruses outside the host: size, structure, com- 
position, organization of virus material; chemical and serological 
analysis; effect of physical and chemical agents (chapters 4-7). 

4. Virus-host interaction: analysis of the simplest and most thor- 
oughly investigated host-virus systems; study of virus reproduction; 
interaction among viruses in common hosts (chapters 8-14). 

5. Viruses in nature: variation, ecology, survival, transmission; virus 
latency and relation of viruses to growth and morphogenesis; tumor 
viruses (chapters 15-17). 

6. Origin and nature of viruses; relation to other biological systems 
(chapter 18). 

7. Rickettsiae-( chapter 19). 


It is customary and reasonable to subdivide viruses, according to 
the major subdivisions of their hosts, into bacterial viruses (bacterio- 
phages), plant viruses, and animal viruses. We must realize, however, 
that even such broad subdivisions may create ambiguities, as in "plant 
viruses" that reproduce in their insect vectors. We must remember 
that, because of their nature, viruses are detected and discovered as 
pathogens, that is, as agents causing abnormalities in some hosts. It 
is therefore logical to list them in terms of their "major host," that is, 
of the host whose manifestations are of the greatest importance to 
maneconomical, medical, or otherwise. In general, each virus will 
have a variety of hosts, more or less related organisms, in which it can 
reproduce. Some of them, often those in which damage caused by 
the virus is slight or absent, are more important in assuring the survival 
and evolutionary success of the virus than the major host which arouses 

CH. 1 Range of Existence of Viruses 9 

man's interest. In tables 1 and 2 we present an extensive though by 
no means complete list of animal and plant viruses. 

There are several important aspects to be considered in taking stock 
of the range of organisms that have been found to be hosts for viruses. 
As far as bacterial viruses are concerned there is hardly a group of 
readily cultivable bacteria for which bacteriophages are not known. 
The bacteria for which no phage has been described ( spirochetes; 
myxobacteria; iron, sulfur, and nitrifying bacteria) present major 
technical problems to bacteriologists. Our knowledge of them is quite 
inadequate. It is likely that every thoroughly investigated bacterial 
group will be found to be the host of some phages. It is interesting 
to note that the host range of phages does not cut across well-estab- 
lished taxonomic boundaries between bacterial groups. Phages active 
on micrococci will not grow on streptococci; phages of enteric bacteria 
do not attack Pseudomonas. Specificity can go far beyond the rather 
flimsy classification boundaries that separate genera and species. It 
reaches down to individual strains or "clones." The explanation for 
this great specificity resides in the fact that phage resistance in bacteria 
is acquired by discrete mutational steps, so that a strain sensitive to a 
phage may give rise to stable mutants resistant to that specific phage. 

Among animal viruses, the only invertebrates in which virus diseases 
have been observed are the insects. These represent, of course, the 
economically most important and therefore scientifically most promi- 
nent group. The study of insect virus diseases was stimulated in 
France in the 19th century by the losses due to diseases of the silk- 
worm, the protagonist of the natural silk industry. Virus diseases of 
insects, especially Lepidoptera and Hymenoptera, are today a most 
important area of virology. 

Among vertebrates, virus diseases have been recognized in fish (carp 
pox, infectious tumors) and in amphibia (virus tumor of the kidney in 
the leopard frog). In birds we find virus diseases of great economic 
importance: Newcastle disease of fowl, laryngotracheitis and many 
others. The main importance of some virus diseases of birds is their 
occasional transmission to man (psittacosis, ornithosis). Certain neo- 
plastic virus diseases of birds, fowl sarcomas and fowl leukemia, have 
a tremendous interest for the virologist because of their role in the 
study of the relation of viruses to tumors. They represent in some 
respects the most thoroughly investigated cases of tumors caused by 

In mammals, virus diseases have been recognized in most domestic 
animals and in several wild ones, particularly in rabbits, whose virus 


The Science of Virology 


" * fl cJ .>^>rt*rt rt , rj *> | rt 

^ S o a o d o fl o s o s dfl^ g 
.2 -5 2 '5 2 '5 2 'S 2 '3 2 '5 2^2^ 



* a . a . a a a -3 a a 1 s *f 

.S^'d <a"g yq " ca 'g c 'g < H <3 13 < g 'S .S 

^ 4- 4J *-> 4J 4-Jj4->^j4jij -^4^3 ^^^'^n^ 

111 11 1111 11 11 IllJfrJ 

dJJ< ^^ JH<a.<JHw - M ^73^3-fST; 

M Ug g,g agaglxg ag ^g^'gg^ 

2 g> g'p '>'?'> |'> I'Pgcaa^ 



Typical Reactions in 
Major Host 

j - 

$ l 1 f Ills !i 

a i | 1 v ^ g M 3 ^S 
.a S ? ? .a .-a 3 1 > .5 
a",, a a -a -a 1 ?J- l.s s 

1|I 1 1 1 11 | sS-H fll 

^j W ^4 1 1 U^ U^ ^^ B w V V . 

g^ g 222 pq u > u^ 




73 ^^ 73 

*rt O -2 5; 

J 8 | | ^ 

"o "3 '3 "3 "S ^ t3 '^ o /^N "o t 

3 1 | l- 2 ^ ^ 3 8 'g g 3 ^ 

d o o^ ** ^ r\r^. r 

oS 2 SUH S u^T UL 


Other Natural 

S ^ *"* 
S S ^ 4.J r73 j 

i 1 iii Ji lit ' 

2 u 2 J^ u 


dd 2 ddd d dd dc 

03 cj Q cd 03 c3 eg oj o3 03c 


' i i .2 T* 

* 4) O c8 ^ *4J 

0.4 1 1 j i,"! 1 I 1 ^ -S 
=JsS>^fl& 2 2 - 

III !!! i 1 ! ! t ] 

1i I l ! ; 

SrtW h^h^PCH CO SB "3 4 

UQ W H M J3 H 

CH. 1 

Range of Existence of Viruses 


*4 |1 M ..g- 9 
V s gis ill ill 3 
la 111 a|l.15 < 8' 

I?.?! 81 ?. S-l^JI-a- 

To c !< ^ - G 3 J2 S C 3 

SS.s.s *> g.2 Tjl.2.2 g-^-g 

cj ** S3 --5 * & W) -M < 

g *s -3 g 8 s : a 3 1 & 

^ -^ ^ ^ ^g^cc^o- 

^ <X) * >> 

2 ..i .. S 

li: .^. T i 

^^Sg^Sfe tS 

.S H S 1 S -5 -| |.S * 

^ i ^-M'^^-^^cS > -g ^J 
3 ojflrt.ep3 *nS C3 

g -B S.S b^0_o| i 
g ^J-S^I^^JgS | 

! 1|1I|H311 | 

HH \J Vj I ( U 

pj W5 

.22 "^ aj 

1 1 

!"5 12 "^ 

r en 

._g ^ T3 

oJ *jj W) ~ .22 

| g en 

'S S.S S 

3 3 .2 * " 

1 3 s ? w-| 1 

'g J a g 1 S J 

- D ^ i; ti 3 e3 

1 *: 111:1 ? 

CX 5-^ ^CC^ g 

b" . .S .g .S 2 I 

.S 2.g .s 1 |\g l.s 

^ C ^,1403 l ( 

Ph PH C/3 C PH ^ 

C/3 O C/2 C/5 1-5 c/3 C/2 




"S "S "1 "1 " 1 

2 -*-> --> -2 *J -M 

o o o cj "5 v o 

03 o3 oj o3 M o3 oj 

63 S3 63 (3 C C 

" "S "S " g "S "S 

O O O O O O 

O O O O L^d O O 


U U U U S U U 

I * a 


1 1 1 1 

1 i -2 2 "^ jg so 1 


"S o !s o 'c c 

-^ Cu o CX O c 


63 63 63 63 63 63 

63 C 63 63 63 d 5 

^5 ^5 ^5 ^5 ^5 ^ 

J3 bS ^3 v3 ^ ^ "S 

& S ^ 

'^ v en 

> 2 en 

x -s O ^ 

PH ^ 

J I b 

| s ^ 1 

<? 8 .2 
ej a 

Ss *-| "g fe 2 li 

u 8 ^3 

S o| g w ^B 
g -, -S 3 | 03 

S g fe S = S .2 HH 

Efl S^os ^H 

A $ gg (5 

1^ fl flj Vs -*' l~ t* 

llll -11 ill 

|a|| |3 ||| 


The Science of Virology 











ical R 




















'< O 


Z 4-) *-> 

U U 


r Nat 
















CH. 1 

Range of Existence of Viruses 


n (_l r-i 

Cg D H -3 

T3 s ^ * 

y o -5 .> 

2 CQ Q -4- 

M - "3 



S -2 c 

3 S 


bC -M 



S ~3 


= .2*3-= 

'53 c S2 o S 
g '^ ,g S 


cij w 



'^S-CS^ S 




u u u 


+J 4^ *J 

Ii 3 3 


o o 
U U 








fi C C C 

,o e o ,o ^o t- 

U U U 6 



e^ q; 

.s .ti .ti .ti "f & 

^3 ^ -S .2 ^ 


S a 






3 -j 22 

v .2 -M 

!H "o '^ 

E i> ' 

I S 






S. s 




The Science of Virology 
Table 2. Representative plant viruses 

Name of Virus 


Abutilon variegation 

Aster yellows 

Bean mosaic 

Cocoa swollen shoot 

Corn streak 

Corn stripe 

Cucumber mosaic diseases 

Elm phloem necrosis 

Pea mosaic 

Peach mosaic 

Peach rosette 

Peach yellows; little peach 

Potato leaf roll 

Potato paracr inkle 

Potato virus X 

Potato virus Y 

Potato yellow dwarf 

Rice stunt disease 

Sugar beet curly top 

Sugar beet mosaic 

Sugar cane Fiji disease 

Sugar cane mosaic 

Tobacco mosaic v 

Tobacco necrosis 

Tobacco ringspot 

Tomato bushy stunt 

Tomato spotted wilt 

Tristeza of citrus 

Tulip break 

Turnip yellow mosaic 

Wheat mosaic 

Wound tumor 

White flies? 

Mechanical; seed; aphids 
Mealy bugs, carried by ants 

Mechanical; seed; aphids; cucumber beetles 



Mechanical (artificial only) 










Mechanical (soil to roots) 




A pi lids 


Flea beetles 

Mechanical (soil to roots) 


* Generally includes name of major host. 

tumors (papilloma, fibroma) have provided an invaluable material foi 
the study of the relation of viruses to neoplasms. Virus diseases ol 
man (over 40 are now known) include such major epidemiologica 
problems as smallpox, yellow fever, poliomyelitis, measles, mumps 
rabies, and various types of encephalitis. 

In the plant kingdom, the absence of reports of virus diseases ir 
lower plants (except bacteria) is notable. Only the flowering plants 

CH. 1 Range of Existence of Viruses 15 

or angiosperms, have yielded viruses. It is difficult to decide whether 
viruses for other plants do not exist or have failed to be detected 
because of limited knowledge. We could reasonably expect that some 
virus diseases of yeast, of molds, or of gymnosperms would have been 
observed if they were at all frequent. 

Among the flowering plants we have most of the plants of economic 
importance. Here viruses probably rank second only to fungi as 
agents of diseases of practical importance. Virus diseases of such field 
crops as potatoes, beans, beets, tobacco, sugar cane, and of fruit crops 
such as peaches are among the major problems of plant pathology. 

Some viruses stand out historically as particularly important in 
general virology, apart from any practical importance, because they 
have provided model cases on which general principles .or methods of 
virology have been worked out. Thus, the study of the chemistry of 
virus particles has progressed mainly through work on some plant 
viruses and bacteriophages. The physicochemical analysis of size and 
shape of viruses has been furthered especially by the study of tobacco 
mosaic virus and tomato bushy stunt virus. The analysis of virus 
reproduction has been carried furthest with bacteriophages and with 
the viruses of the influenza group. The study of tumor causation by 
viruses has derived much of its impetus from the discovery of virus- 
induced chicken sarcomas. 

Yet it is essential to keep constantly in mind in the study of virology 
that no conclusion based on the study of one virus can a priori be 
generalized as valid for any other virus. In view of the presumed 
heterogeneity of the objects that we call viruses, the greatest caution 
must be exerted in attributing to any one virus a property observed 
in another. This does not mean that we should consider each virus as 
a completely distinct entity, unrelated to any others. Indeed, we 
recognize certain groups of viruses (those of the psittacosis-lympho- 
granuloma group, for example, or certain bacteriophages) as consti- 
tuting taxonomic groups as valid as any found in higher organisms. 
We must simply keep in mind that viruses as a group may include 
things as different in their structure, chemical composition, and bio- 
logical activities as any odd collection of biological materials selected 
on the basis of arbitrary criteria, for example, all animals shorter than 
an inch or all plants with pink flowers. 

16 The Science of Virology 


The common nomenclature for animal and plant viruses consists in 
using the name of the disease produced in the major host followed by 
the word "virus." Bacteriophages are named by code symbols (gen- 
erally letters followed by numbers, as Tl, C16, S13) derived from more 
or less accidental laboratory customs; the symbol is sometimes pre- 
ceded by the Greek letter <f>. This system of nomenclature, which is 
not a system of classification and which frankly admits its own em- 
pirical, nontaxonomic basis, has worked quite adequately for animal 
viruses and for bacteriophages, but not for plant viruses. The need 
and quest for a taxonomic classification, indicating natural relation- 
ships, has been an ever-present problem for plant pathologists. The 
difference may be due partly to a difference in scientific training. 
Plant pathologists are generally trained as biologists, whereas animal 
pathologists are more often physicians or veterinarians. Moreover, the 
natural relationship among plant viruses is of some practical signifi- 
cance, because cross-protection between related viruses infecting the 
same plant provides a potentially important defense mechanism in 
plant virus diseases. The increasing realization of the role of reservoir 
hosts in the epidemiology of virus diseases, both of animals and of 
plants, is also increasing the requirement for a characterization of 
viruses independent of their host range and of the pathological re- 
actions of the hosts. 1 

The shortcomings of all systems of classification of viruses reflect, of 
course, our present inadequate knowledge of the nature and origin 
of these entities. This makes any assessment of the degrees of rela- 
tionship among the viruses highly problematic. On a strictly prag- 
matic basis, a provisional classification can be made using any group 
of virus properties that are sufficiently stable and distinctive. Most of 
our knowledge of viruses, however, concerns the effects they produce 
on their hosts, effects which may reflect the properties of the hosts more 
than those of the viruses. 

Plant pathologists have repeatedly proposed classifications or systems 
of nomenclature of some practical value. In some cases, the grouping 
criteria were the properties of virus materials in vitro. In others, the 
criterion was the plant host in which a virus was first described (see 
43). Clearly, the latter method is of little significance as to virus rela- 

1 The problems of virus classification were discussed from many viewpoints at a 
symposium held in New York in 1952 ( 121a). 

CH. 1 Classification of Viruses 17 

tionship, since most viruses attack many hosts. Holmes (341) has 
proposed a Latin binomial system, similar to that used for plants and 
animals, and has extended it to include all viruses. These would form 
an order, Virales, with three suborders, Phagineae (bacteriophages), 
Phytophagineae (plant viruses), and Zoophagineafc (animal viruses)^ 
further subdivided into families, genera, and species. The properties 
used in grouping and separating viruses into groups are mainly symp- 
toms produced and mode of transmission. A major weakness of the 
Holmes system is the utter disregard for the morphological properties 
of virus particles as a criterion of classification. Thus, viruses such as 
the foot-and-mouth agent with its tiny particles, and the large-particled, 
well-characterized viruses of the vaccinia group are placed together in 
one genus because they attack mucous membranes. 

In the mjnds of many virologists, attempts to set up a complete or 
even a detailed system of classification are not only premature but also 
misleading. The adoption of Latin binomials suggests, by analogy 
with plant and animal taxonomy, a similarity of evolutionary pattern 
and of taxonomic hierarchy between viruses and higher organisms, 
which is wholly unestablished. The attempt to classify viruses about 
which little is known, except a description of the disease produced in 
one host, leads to an unjustified feeling of knowledge and full under- 
standing. It would seem better to admit that we know a few things 
about a few viruses, whereas many viruses probably exist about which 
we know nothing. Then we could describe the best-known viruses in 
terms of their own properties, compare them among themselves, and 
organize a few main groups having as prototypes those well-known 
viruses that are clearly distinct from one another. One aid in this 
process would be the knowledge that virus variation occurs in the form 
of discrete mutational steps (see chapter 15). Properties known to 
be changed by a single mutation, such as host range and tissue affinity, 
would be taxonomically less valuable than those found to be more 
stable, such as morphology or serological specificity. 

Some plant virologists (43) and some bacteriophage workers (9) 
have reached similar conclusions on taxonomic questions. These in- 
vestigators consider as the primary criteria of natural relationship the 
similarity of particle size and shape and the possession of common 
antigens revealed by serological cross-reactions. For example, two 
viruses such as the tobacco mosaic agent, with rod-shaped particles of 
characteristic size and composition, and the tobacco necrosis virus, 
with spherical particles and a tendency to form crystals, can be used 
as prototypes for two groups of viruses, which include strains differing 

18 The Science of Virology 

from the prototype by one or a few mutational steps. At the level of 
larger groups these two viruses could, respectively, serve as prototypes: 
the first for plant viruses with rod-shaped, fairly rigid particles; the 
second for spherical, regularly crystallizing plant viruses. 

Similarly, coli-phages T2, T4, T6, and C16, which are indistinguish- 
able in morphology, give serological cross-reactions and are capable of 
genetic recombination in mixed-infected hosts (see chapter 8), form a 
natural group, probably equivalent to a species. The various types 
may be considered as varieties or incipient species ( 9 ) . As for defining 
larger groups, there seems to be little basis on which to proceed. 
Although phages with very different bacterial hosts (for example, 
micrococci or coli-dysentery organisms) do not seem to adapt from 
one such host to the other, they are often morphologically more similar 
(although serologically unrelated) than phages with common hosts. 
This leaves open the question of host range versus morphology as a 
criterion of relationship. In plant viruses, cross-protection reactions are 
well correlated with serological ones and permit additional groupings. 

Criteria analogous to those useful in grouping plant viruses and 
phages can sometimes be applied to animal viruses. For example, the 
viruses of the smallpox-vaccinia-ectromelia group are a morphologically 
homogeneous group with antigenic components and other properties 
in common. It has been suggested (259) that this group and other 
viruses, such as fowl pox and molluscum contagiosum, be placed in one 
genus (Borreliota)^ all the members possessing fairly large particles 
and giving typical intracytoplasmic inclusions. 

Another homogeneous group is the so-called psittacosis-lympho- 
granuloma-pneumonitis group, consisting of several viruses with large 
particles, which give serological cross-reactions and undergo similar 
developmental transformations inside their host cells. The infections 
produced by these viruses are susceptible to therapy by antibiotics. 
These viruses have been named "Chlamydozoaceae" and have re- 
ceived rating as a family of the order Rickettsiales in the 1948 edition 
of Sergey's Manual of Determinative Bacteriology (Baltimore, Wil- 
liams & Wilkins). 

Another natural group may be that of the influenza-mumps-New- 
castle disease viruses, which have similar particles and common 
enzymatic activity on some surface component of red blood cells and 
on other substrates. Serological cross-reactions are not present among 
all members of this group, however. Other groups with reasonably 
good taxonomic standing may, for example, be the rod-shaped insect 
viruses producing crystalline intranuclear inclusions containing the 

CH. 1 Classification of Viruses 19 

virus, and the equally rod-shaped viruses of insects whose particles are 
enclosed in characteristic ellipsoidal capsules (68a). 

In summary, a wisely conservative position appears to be that the 
more properties of viruses we learn to analyze, the more meaningful 
will be the relations we are able to establish. With bacteriophages, 
where genetic studies have progressed quite far, we may be approach- 
ing a point where genetic facts will be used to provide a rationale for 
taxonomic principles (9). For example, we are sometimes able to 
invoke as a criterion of classification not only the possession of a com- 
mon character but also the possession of common mutability, which 
strongly suggests the possession of similar genetic determinants. This 
is true of the coli-phages T2, T4, T6, and C16, which undergo similar 
mutations. Here, the possession of common genetic determinants has 
been shown by tests of genetic recombination. In viruses of the 
tobacco mosaic group parallel mutability has also been described. 


Detection and Identification 
of Viruses 

In studying viruses, we must first consider ways of recognizing that 
they are present. The essential thing to look for is some abnormal host 
manifestation, generally macroscopic. In their extracellular, free state 
viruses are inert and do not act on their environment. Because of 
their small size they cannot easily be recognized. Although in prin- 
ciple it should be possible to detect and recognize a virus from its 
morphology, particularly when an electron microscope is used, the 
practical value of such a method is small. Only when the presence 
of a virus is suspected because of some abnormal host manifestation 
does the investigator have recourse to attempts to locate a virus by 

[Host reaction to a virus is recognized by comparing infected and 
normal host. No general rule can be given. Such recognition depends 
on our familiarity with the symptoms of the various virus diseases. 
This is what we might call the "art" of virology. At the basis of any 
scientific construction is a whole body of information, which the in- 
dividual worker can learn only by experience and by accumulation of 
information. This we call the "art" of the scientist. For example, the 
background of all chemical generalizations is the acquaintance of the 
investigator with large numbers of chemical facts. These permit him 
to devise chemical tests and provide a basis for him to interpret the 
properties of chemical compounds. 

\ For the practical purpose of interpreting the virologist's detection 
methods we may subdivide virus manifestations into two main cate- 
gories: local reactions, and generalized or systemic reactions.; 


CH. 2 Local Reactions 21 


(jThese are manifestations that involve only a limited number of host 
cells localized in one area. The limitation in size of the affected area 
is due either to a localization of the virus in a few cells or to the fact 
that only a few cells show the effect of the virus' presence, although the 
virus may be present beyond the limit of the recognizable lesion. A 
virus may remain circumscribed in a local area because of its limited 
ability to spread, because of a limited number of susceptible host cells, 
or because of a balance between host defenses and virus invasiveness] 

Bacteriophage. The local lesion of bacteriophage is tailed a plaque 
arid is an area of lysis or dissolution in a bacterial layer growing on a 
solid medium. The plaques are often quite sharp and round, and may 
be surrounded by haloes of incomplete lysis. Formation of discrete 
plaques is due to the fact that bacteriophage reproduction can only 
occur in growing bacterial cells. On a nutrient medium there is a race 
between two processes: bacterial growth and bacteriophage growth. 
When bacterial growth stops, lysis by bacteriophage also stops or at 
least proceeds at a very slow pace, so that the area of lysis remains 
localized. At the same time, the bacteriophage particles also diffuse 
through the agar and in the liquid film at its surface. Before growth 
stops, they may have reached bacteria at a certain distance from the 
original phage localization. As a consequence, smaller phages, which 
diffuse more rapidly, give relatively larger plaques on a given bac- 
terium than larger phages do. 

The appearance of the plaque is characteristic for each phage-host 
system (figure 1), and reflects the kinetics of bacteriophage growth 
and of bacterial growth, and the size of the bacteriophage. Size and 
morphology of bacteriophage plaques also depend on the composition 
and consistency of the nutrient medium, which affects host growth, 
phage growth, or both, 

Plant viruses. Local lesions are generally detectable on the surfaces 
of leaves. ' The most interesting type is the necrotic lesion, a black or 
brown spot that may vary, when fully developed, from 1 to 10 mm in 
diameter, and which appears at the point of entry of some viruses, 
for example, of tobacco mosaic virus in Nicotiana glutinosa and in 
other plants. The virus enters a cell at a point where some mechanical 
damage has been produced on the leaf, and then spreads from cell to 
cell. A necrotic lesion results from rapid death of a group of infected 
cells, with consequent localization of the virus in a small area. The 

22 Detection and Identification of Viruses 

nggrptic reaction is_generally caused byjiyperjgnsitiyity; the cells die 
so fast that virus spread is limitecT Itactually represents a defense 
'mechanism bbTfrf or the individual plant, which can localize the infec- 
tion and survive, and for the crop, since virus localized in jiecrotic 
lesion jias fewer chances of spreading from plant to plant than a virus 
that gives a generalized^ infection. 

Figure 1. Plaques produced by different coli-phages on bacterial layers of a 
common host, Escherichia coli, strain B, growing on agar. Reduced to about % 
size. From: Demerec and Fano (177). Courtesy Dr. M. Demerec, Carnegie 
Institution, Cold Spring Harbor, New York. 

Necrotic ' lesions have shapes and sizes characteristic for a given 
virus-host system (see figure 2). Necrotic lesions are particularly im- 
portant (333) because they provide the best method for titration of 
those plant viruses that produce them (see chapter 3). 

The presence or absence of necrotic lesions at the point of entry of 
the virus has been traced to specific genes of the host. The necrotic 
reaction to tobacco mosaic virus in species of the genus Nicotiana 
(337) is due to a dominant gene^ which determines hygersensitivity, 
whereas the recessive allele causes a systemic mosaic response. 

A well-analyzed example is that of tobacco mosaic virus in the pepper 
(339). The type of response depends on three allelic forms of the same 


(c) (d) 

Figure 2. Local necrotic lesions produced by plant viruses on plant leaves, 
(a) Tobacco mosaic virus on Nicotiana glutinosa; (b) tomato mosaic virus on 
N. glutinosa; (c) tomato bushy stunt virus on N. glutinosa; (d) tobacco ringspot 
virus on N. tabacum. Courtesy Dr, K. M. Smith, Molteno Institute, University of 


24 Detection and Identification of Viruses 

gene: one allele, L, determines a localized necrotic response; another, 
l\ produces, when homozygous, a delayed necrotic response followed 
by abscission of the leaf and recovery of most plants; the third one, /, 
causes generalized chlorosis. 

Sometimes, although the response at the point of entry is necrotic, 
localization does not result and the virus continues to spread. This 
is true, for example, of some potato viruses. Potato virus X in some 
varieties of potato elicits a necrotic response; in other varieties it gives 
generalized mottling or even symptomless infections. Even with local 
necrotic response, the necrotic reaction may become generalized; the 
localized infection, when present, may be due to virus inactivation in 
the plant rather than to hypersensitivity. The genetic basis of the 
responses of potato varieties to virus X has been partially clarified 
(356a; 629). 

Another type of localized response is the so-called starch-iodine 
lesion, which consists of local areas of altered starch metabolism around 
the points of entry of the virus (335). To demonstrate this type of 
lesion, an infected leaf is treated with alcohol to remove chlorophyll 
and then with iodine to stain starch. The lesions appear as areas 
lighter than the rest of the leaf, if the leaf had recently been exposed 
to light, because of reduced synthesis of starch in the infected areas. 
Or, they may appear as darker areas, if after exposure to light the leaf 
had been allowed to spend several hours in darkness, presumably be- 
cause the starch moves away less freely from the damaged areas, due 
to hindrance to food transport. 

Animal viruses. Recognizable local lesions are formed either at the 
primary point of entry or by secondary localization of a circulating 
virus. They generally consist of cell destruction and cell proliferation, 
with the complication that .inflammatory reactions are always present 
in animal tissues. An example of a primary localization of infection is 
that of infectious warts in man. A variety of warts or papillomas, flat 
or pointed, may result from the entry of the virus, depending on the 
point of entry and possibly on the virus strain. As examples of sec- 
ondary localization of viruses producing recognizable local lesions we 
may mention the exanthemata that are the major symptoms of smallpox, 
measles, and chicken pox. These typical rashes are caused by the 
localization of the virus in the skin chorion, after which the cells of the 
basal layer of the epithelium are invaded and damaged and blisters 
are formed in the thickness of the epidermis. In some cases, for 

CH. 2 Local Reactions 25 

example, in the Koplik spots that appear on the oral mucosa in the 
early stages of measles, the localization of the virus is in the sub- 
mucosal glands. 

In special instances we may observe local lesions on layers of cells 
that are more or less uniformly susceptible. An example is the infec- 
tion of the chorioallantoic membrane of the chick embryo (figure 3; 
see chapter 12). A virus suspension is deposited on the surface of the 

(a) (b) 

Figure 3. Chorioallantoic membrane lesions, (a) Myxoma virus; (b) Murray 
Valley encephalitis virus. Courtesy Dr. F. M. Burnet, Walter and Eliza Hall 
Institute, Melbourne. 

membrane and spreads over it. Here and there, the virus gains access 
to a cell; this leads to the formation of a lesion. At first the lesions 
are discrete, but later become confluent if the infection proceeds long 
enough. The conditions are somewhat similar to those of a bacterial 
layer seeded with bacteriophage, although less well understood. Local- 
ized lesions can also be observed in layers of cells cultivated in vitro 
(figure 4; see chapter 11). 

Sometimes, localized lesions even in deep tissues may be utilized in 
the diagnosis of a virus, because of either their localization or their 
appearance. Poliomyelitis virus produces areas of cell destruction in 
the anterior horns of the spinal cord, where motor cells are localized, 
with a resulting muscular paralysis. Rabies virus causes typical altera- 
tions in the cells of the Ammon's horn in the brain. These cells con- 
tain the characteristic intracellular inclusions of rabies (Negri bodies). 
Many similar examples could be cited from the pathology of animal 
virus diseases. 


Detection and Identification of Viruses 


There is no clear-cut distinction between generalized and local 
reactions, since all host manifestations in a virus infection depend 
ultimately on the invasion of sensitive cells by virus. Nevertheless, the 

Figure 4. Plaques produced by Western equine encephalomyelitis virus on a 
layer of chick embryo fibroblasts in tissue culture. Courtesy Dr. R. Dulbecco, 
California Institute of Technology, Pasadena. 

recognizable symptoms of infection are often more diffuse than the 
appearance of localized areas of damage. For example, we may con- 
sider the mass lysis of a bacterial culture by bacteriophage as a gen- 
eralized infection, as opposed to the localized formation of plaques on 
a layer of bacteria. In both cases we observe lysis of large numbers of 
cells, but in the first there is no spatial limitation to the spread of the 
virus, such as is caused on a plate by the immobilization of the bacterial 
cells and by the limited means of dispersion of the virus. 

Plant viruses. Generalized or systemic infections produce symptoms 
that go under a variety of names: chlorosis, mosaic, streak, yellows. 

CH. 2 

Generalized Reactions 


Figure 5. Aster plants. The plant on the right is normal. The plant on the 
left is infected with aster yellows virus. Courtesy Dr. K. Maramorosch, Rocke- 
feller Institute, New York. 

Figure 6. Leaves of Nicotiana tabacum. Left: normal. Center: infected with 
tobacco mosaic virus. Right: infected with the same virus, with a bright spot of 
yellow mosaic mutant. Courtesy Dr. L. O. Kunkel, Rockefeller Institute, New 


Detection and Identification of Viruses 

ringspot, leaf roll, describing the type of deformation or discoloration 
that appears on the leaves of the infected plants (figures 5 and 6). 
Discoloration of leaves may be localized in ringlike areas (ringspot), 
in streaks, in irregular patches (mosaic), or it may appear as a uni- 
form chlorosis or as a yellowing, often more pronounced along the 

Figure 7. Tumors produced by wound tumor virus on the roots of sweet clover. 
Courtesy Dr. L. M. Black, University of Illinois, Urbana. 

veins. Most of these changes reflect disturbances in the photosyntnetrc 
function or in the transport of food, as we have already mentioned for 
the starch-iodine lesions. To describe the specific alteration of the 
morphology or habit of a plant affected by a virus, plant pathologists 
have traditionally employed picturesque terms such as sugar beet curly 
top or potato witches' broom. 

Some symptoms result from cell proliferation in virus-infected tissues, 
which lead to production of "enations," frilly or compact growths de- 
rived from the vein areas of the leaves. Lateral buds may proliferate 
to produce the symptoms of witches' broom, One virus produces 

CH - 2 Generalized Reactions 29 

multiple tumors throughout the infected plants; the tumors arise mainly 
in areas where the plants are wounded (83; figure 7). 

Flowers may show symptoms of virus infection in the form of varie- 
gated colors or "breaking." Tulip break, a condition which has been 
known for several hundred years and which produces a highly appre- 
ciated flower appearance, is due to virus. 

The symptoms of a generalized plant disease are usually visible on 
the new leaves formed after the infection has become systemic; old, 
nongrowing leaves show symptoms only if infected directly. Generally 
the newly growing parts of the plant show the greatest damage. On 
tfie other hand, apparent recovery occasionally occurs and the new 
growth is either free or almost free of symptoms, although it generally 
still contains virus. The type of manifestation, moreover, depends 
on the host. A virus may be very destructive for one host and cause a 
mild or latent disease in another. Different varieties of potato, for 
example, differ widely in their responses to potato viruses. 

Animal viruses. Detection of animal viruses is even more complex 
than detection of plant viruses. This is due to the greater differen- 
tiation of tissues and organs, to the presence of several mechanisms of 
virus transport, to serological phenomena of immunity, and to sec- 
ondary inflammatory processes. We may mention, as examples, the 
symptoms of mumps, in which the virus localizes mainly in the parotid 
gland, but occasionally produces an inflammation of the genital organs; 
and those of yellow fever, in which the virus may be widespread in 
many different organs and tissues. The general symptoms can always 
be traced to damaged and altered cell function. For example, the 
jaundice that accompanies yellow fever is due to the damage to the 
liver cells. Detailed descriptions of the manifestations of virus dis- 
eases of man and animals are found in special treatises (414; 655). 

An interesting aspect of the general manifestation of animal virus 

leases is the rather precise incubation period between infection and 
appearance of symptoms. For example, the incubation for measles is 
generally 10-12 days, for mumps 18-21 days, for smallpox around 12 
days. This suggests that a fairly definite sequence of events must take 
place before the appearance of symptoms. The incubation period may 
include the time needed for the virus to penetrate (and possibly to 
reproduce at the point of entry), to circulate, to localize itself in sus- 
ceptible cells, and to reproduce to the extent needed for the mani- 
festations to appear. 

We often encounter silent infections with animal viruses. The virus 
reproduces without giving visible symptoms. In many cases, however, 

30 Detection and Identification of Viruses 

the presence of the virus can be traced by the presence of circulating 
antibodies in the blood. This sign of infection is useful in evaluating 
the degree of immunity and in following the spread of a virus through 
a host population. The serological detection of virus infection forms 
an important phase of the epidemiology of virus diseases. The pres- 
ence of antibody is not proof of the actual presence of the virus in the 
organism, because a virus disease may give rise to a lasting immunity 
and continuous production of antibody even after the virus has ap- 
parently disappeared from the organism. This phenomenon is relevant 
for the general problem of the mechanism of antibody production. It 
may be that the virus does not disappear completely, but only vanishes 
as an infectious material and remains present in some nonrecoverable 
form in the cells that produce antibodies. 


In all cases where viruses are identified from the symptoms they 
produce, it is important to define the situation in which we observe 
such symptoms. We have already mentioned several examples of the 
influence of the host species on the type of manifestation of a virus. 
A bacteriophage may completely dissolve the cells of a certain bac- 
terium and may be carried without obvious manifestations within the 
cells of another bacterium ("lysogenic strain") (120a; 453). The 
pseudorabies virus causes a fatal disease in cattle but a mild, almost 
nonsymptomatic, although very contagious, infection in swine (602). 
As we shall discuss in greater detail in chapter 16, whenever a virus 
is very destructive for a certain host we have reason to suspect that it 
has some other host in which it produces a mild disease, because only 
such a host-virus compromise can provide a virus with opportunity for 
successful survival in nature. Too complete a destruction of the 
host cuts the ground from under the virus and tends to lead to its 

The genetic basis of virus susceptibility in different hosts is em- 
phasized in clear-cut fashion by such observations as that on a strain 
of mice that owes its 10035 resistance to strain 17D of yellow fever 
to a single dominant gene. In the brain of the resistant mice the virus 
fails to multiply (572). 

Another important factor is the age of the host. Young plants are 
generally more susceptible to plant viruses. In animals, the symptoms 
produced by a given virus in the same host often differ with the 

CH. 2 Influence of Host and Environment 31 

host's age. The tumor virus that produces the Rous sarcoma in adult 
chickens causes an inflammatory, nonproliferating disease in the newly 
hatched chick (198). Such differences in host reaction reflect the 
changes in the physiology of the host cell that take place during 
development and differentiation of various tissues, and also depend 
on the appearance or disappearance of certain specific defense mech- 
anisms in the course of aging (for example, low antibody production 
in the newborn ) . 

Symptoms also depend on the route of entry of a virus. For example, 
influenza virus entering a mouse by the usual route, through the 
mucosa of the upper respiratory tract, causes typical pneumonia; if 
injected in the brain, it may give a toxic syndrome (298). The toxic 
manifestations are apparently caused by an abortive reproduction of 
the virus in the cells of the central nervous system (582). 

Other factors that influence symptomatology are the environmental 
conditions prevailing during the infection period. We have already 
mentioned the influence of light on the appearance of starch-iodine 
lesions at the points of entry of certain plant viruses. Season, soil 
constituents, and many other factors also affect the manifestations of 
plant virus diseases. Nutritional effects on the symptoms for animal 
viruses and also on the susceptibility to animal viruses have been 
noticed; greater resistance seems sometimes to be associated with 
some nutritional deficiencies (134). This suggests either that the 
shortage of an essential growth factor limits virus growth, if the 
requirements of virus-infected cells are greater than those of un- 
infected cells; or that the cells of growth-factor-deficient animals are 
so altered physiologically as to support less virus growth or to resist 

Temperature effects on virus disease have often been observed, 
especially in plants. Some viruses are unable to grow at temperatures 
above a certain level. An infected plant may become completely 
sterilized by a few hours or days of immersion in a water bath at a 
temperature as low as 35 C. This is so for several diseases of the 
peach tree (388). Temperature may be the reason why such diseases 
are infrequent or absent in the southern part of the United States, 
where the summer temperature is often higher than the virus can 
stand. The symptoms themselves may be modified by temperature. 
Tobacco mosaic virus at temperatures around 35 C causes the ap- 
pearance of milder symptoms, due to the selective proliferation at this 
higher temperature of a less virulent variant of the virus (338). 

32 Detection and Identification of Viruses 


In studying a new virus disease and in attempting to identify its 
causative agent, we must meet requirements comparable to Koch's 
postulates for bacterial infections (555). We must prove that an 
agent is regularly isolated from cases of the disease, that it can be 
propagated in a suitable host, that it is filtrable (submicroscopic), 
that it reproduces only in living cells, and that, reintroduced into the 
original host or into a sufficiently similar one, it gives rise to the orig- 
inal syndrome. 

The techniques for successful artificial transmission of a virus to a 
host are part of the art of virology. Success in a specific case depends 
mainly on the investigator's knowledge of the properties of the virus 
he is dealing with. The main problems are: the choice of the proper 
host and of the proper route of inoculation; a sufficient amount of 
virus in the inoculum; and the control of the environmental conditions. 
The essential requirement is to bring the virus, or to cause it to be 
transported, to susceptible cells where it can reproduce. In general; 
the outer surface of most higher plants and animals consists of virus- 
proof barriers, so that the virus must be introduced either through this 
barrier or into some openings through which it can come directly in 
contact with susceptible cells. 

The simplest case* is again that of bacteriophage. In a suitable 
medium, containing the necessary activating factors and ionic concen- 
trations (see chapter 8) a phage particle infects a bacterium by simple 
collision, with a collision efficiency close to 100%. The presence of 
bacterial secretions such as capsules or slime may sometimes hinder 
successful contact. 

Plant viruses. Several means of transmission of plant viruses may 
be listed. Some means apply to most viruses, others only to a few. 
The simplest is mechanical transmission, practiced by rubbing the 
surfaces of leaves with an applicator ( a finger, or a piece of cotton or 
cheesecloth) soaked in the virus suspension. Rubbing causes local 
abrasions, through which the virus can penetrate. Some viruses have 
been transmitted only by such mechanical methods or by grafting. 
A typical example is potato virus X, for which no insect vector is 
known. Many plant viruses cannot be transmitted mechanically. 

The number of points of virus entry produced by rubbing a leaf may 
vary greatly, depending on several factors difficult to control (pressure, 
abrasive power of the applicator). By careful standardization of tech- 

CH. 2 Artificial Transmission of Viruses 33 

nique, however, astonishingly reproducible results can be obtained. 
The points of entry may be broken hair cells or wounded epidermal 

Mechanical transmission, when it is the only method of natural 
transmission of a virus, requires that the virus be widespread in a 
plant population, so that there are frequent chances for virus to be 
carried mechanically from one plant to another. For example, tobacco 
mosaic virus is highly contagious and can be introduced into a field by 
a person whose hands are contaminated with virus, for example, from 
a cigarette containing infected tobacco leaf. Potato virus X can be 
transmitted from plant to plant by the rubbing of leaves blown together 
by wind. 

Insect transmission is the natural method of transmission for a great 
many plant viruses. Artificially, it is utilized by causing insects to 
feed on infected plants within cages, then transferring them to cages 
with uninfected plants. In some instances the insects have been caused 
to feed on cell-free extracts containing virus. An ingenious method 
(59), suitable for the study of the leaf hopper vector of curly top virus, 
Circulifer tenellus, consists of introducing the leafhoppers into a small 
cage consisting of a glass cylinder over one end of which a thin mem- 
brane is stretched. Droplets of the virus-containing extracts are placed 
on the outer side of the membrane. The insect feeds upon them by 
inserting its stylets through the membrane, and can be induced to feed 
for definite lengths of time by carefully regulating the illumination 
and other conditions. Thus, the leafhoppers are fed various amounts 
or concentrations of virus suspension, after which their ability to trans- 
mit the virus is investigated by placing them on healthy plants. It is 
also possible to inoculate viruses directly into the insect body (632). 
Many viruses that in nature are transmitted only by insects can be 
studied in cell-free extracts by injecting the extracts into insects, which 
thereby become sources of infection. 

A method of transmission widely applicable to plant viruses is the 
graft method. Grafting of shoots or buds from an infected onto a 
healthy plant causes the passage of virus from the scion into the healthy 
plant, and is followed by the manifestations of systemic infection. The 
great majority of plant viruses can be transmitted by this method. The 
method has not been useful for viruses restricted to plants like corn 
that are difficult or impossible to graft or for viruses that produce only 
local lesions in most hosts, like tobacco necrosis virus. By grafting, 
viruses can also be transmitted to new hosts, provided of course that 
successful intergrafting is possible. 

34 Detection and Identification of Viruses 

A refinement consists in causing a parasitic plant called dodder 
(Cuscuta sp.) to parasitize an infected plant. The special roots 
(haustoria) of the dodder penetrate the infected cells, and the virus 
spreads to the dodder. Infected dodder stems, when later allowed 
to parasitize healthy plants, will often transmit the virus (61; 363). By 
this method it is possible to transmit several viruses from species to 
species, even when these are not intergraf table (391). 

Some theoretical problems arise with viruses -that have been trans- 
mitted only by grafting. For example, a condition known as variega- 
tion in Abutilon, an ornamental plant, was for a long time transmitted 
by graft only (although transmission by an insect vector has now been 
obtained; 505). According to our definition of virus, such a condition 
could be considered a virus disease only because of the similarity of 
its symptoms with those of other virus diseases. An interesting case is 
that of the King Edward potato, which carries the paracrinkle virus 
in a completely nonsymptomatic state (574). Grafting of the King 
Edward potato onto potato plants of other varieties causes them to 
show a typical disease, transmissible from plant to plant by graft. The 
paracrinkle virus has recently been isolated and inoculated mechani- 
cally; its properties resemble those of several other potato viruses (52). 
Interesting questions have been raised, however, as to the origin of 
such a virus. Does it arise de novo in the King Edward potato by a 
change in some nownal cell protein? Or does it enter this plant 
variety and remain latent, reproducing for innumerable generations 
without causing symptoms? Seed transmission of such a virus is not 
necessary, because potatoes are propagated from tubers. These prob- 
lems will be discussed in detail in chapter 18. 

Animal viruses. The simplest method of inoculation, when feasible, 
is the instillation of virus-containing material into cavities lined with 
susceptible cells, for example, in the extraembryonal membranes of the 
chick embryo or on cells growing in tissue cultures. Instillation in the 
respiratory tract is often successful for viruses that proliferate on the 
epithelium of the lung. It imitates the probable route of natural in- 
fection with viruses that are supposedly transmitted by droplet infec- 
tion, such as the viruses of common cold, influenza, and probably 
measles and mumps. It may be possible to infect animals by feeding 
such viruses as that of infantile paralysis, which are supposed to pro- 
liferate in the intestine. The dropping of virus suspension on scarified 
skin as a method of inoculation is commonly practiced with vaccinia 
virus in vaccination against smallpox. A similar method is the inocu- 
lation of viruses onto the scarified cornea of rabbits and other animals. 

CH. 2 New Hosts for Viruses 35 

Injection into the host's tissues is the most general method of inocu- 
lation; the virus may be injected either directly into the susceptible 
tissue or into a region from which it will spread to susceptible cells. 
Some viruses spread through the circulatory system, others along the 
nerve fibers. Many viruses may reach susceptible tissues if introduced 

Transmission by arthropod vectors has been used experimentally, 
particularly in the study of the t role of insects in the natural transmis- 
sion of a disease. The experiments of Walter Reed and his col- 
laborators in proving that yellow fever is transmitted by a mosquito 
have become classic in the history of virology (548). Insect trans- 
mission is seldom the only possible inoculation procedure, however, 
since animal virus diseases transmitted by insects can also be trans- 
mitted by direct inoculation of infectious material (for example, a 
patient's blood) into a susceptible animal, without need for a cycle 
including an obligate passage through the vector's body. 

Some of the cultivation procedures for viruses represent true enrich- 
ment cultures, by which large amounts of a virus are obtained in a 
suitable host. Sometimes enrichment procedures are necessary as an 
intermediate stage between isolation of a virus and its diagnosis. This 
is the purpose of most "first-isolation" techniques that have been pro- 
posed. For example, in attempting the isolation of viruses from cases 
of influenza or other respiratory diseases in man it has been common 
practice to inoculate throat washings (individual or pooled) into chick 
embryos or into the lungs of mice. After suitable intervals of time 
the inoculated tissues are extracted and tested for presence of virus, 
either by successive reinoculations or by attempts to detect the virus 
through some of its properties. Another example is that of herpes 
virus, which produces typical intracellular inclusions in the cells of 
the rabbit cornea. Material taken from a patient must often be passed 
through the rabbit testis before sufficient amounts of virus are obtained 
for testing by corneal inoculation (185). 


The importance of the correct choice of host in virus transmission 
has already been mentioned. It is often the key to successful work 
with a virus. Because of the laboratory conditions, it is seldom pos- 
sible to cultivate a virus in the host from which it has been isolated. 
This is obvious where human diseases are concerned and is also true 
in many other cases. The search for new hosts or, as it has been often 

36 Detection and Identification of Viruses 

called, the adaptation of a virus to new hosts, is an important phase 
of virology. 

With plant viruses, success is often obtained by grafting, by the use 
of parasitic dodder, or by choosing -the proper insect vector. The 
results are often quite interesting. For example, it was suspected for 
a number of reasons that the virus causing witches' broom in potato 
and the virus of aster yellows might be identical. In the absence of 
intergrafting, dodder was used to transmit them to common hosts. 
On these hosts it was possible to show that the viruses differed in 
regard to their insect vectors and their host range ( 391 ) . 

Transmission of animal viruses to a new host can often be accom- 
plished by successive transfers, often blind transfers, in the sense that 
the host does not show signs of the presence of the virus or that at the 
beginning the virus does not grow in sufficient amounts to be detected. 
After several transfers in the new host, adaptation often improves and 
the virus may then grow quite well. For example, many viruses can 
be made to grow in some of the cavities of the chick embryo by suc- 
cessive transfers. 

Caution is necessary, however, since in the course of its adaptation 
to new hosts some of the properties of a virus may be changed. It is 
now generally accepted that viruses undergo mutations and that the 
ability to attack a new host may be present only in a few mutant 
individuals out of a large virus population (see chapter 15). In the 
new host only the mutants will proliferate, and they will establish a 
new virus line adapted to the new host. In the course of mutation, 
however, some of the properties of the original virus may be changed, 
even including pathogenicity for the original host. The virus worker, 
therefore, must be careful in drawing conclusions, for example, as to 
the nature of the properties of a human virus from the study of its 
egg-adapted (or mouse-adapted or ferret-adapted) strains. 


Another important caution in the artificial transmission of viruses to 
new hosts is the possibility of encountering latent viruses, that is, 
viruses which are carried without symptoms, but which may give rise 
to pathological manifestations after successive transfers. In attempting 
to isolate a virus by successive transmissions of materials through 
several animals or plants, virus workers have often encountered a 
different and completely new virus, bearing no relation to the original 
disease. Some examples will illustrate this point. 

CH. 2 Latent Viruses 37 

The parasitic dodder used to transmit certain plant viruses may 
itself carry the latent dodder mosaic virus, which is pathogenic for 
other plants (62). Virus III was discovered in rabbits (558) in the 
course of attempts to isolate the chicken pox virus by successive trans- 
fers of extracts from the cutaneous vesicles of human patients into 
rabbit testis. A virus was present in the testes of many rabbits of 
the colony studied, and upon unnatural transfer it revealed itself by 
giving typical symptoms. The virus could also be isolated from a 
similar series of transfers initiated without the original injection of 
material from chicken pox patients. 

If lung extracts prepared from mice of various colonies are serially 
inoculated from animal to animal, after a number of transfers (from 
3 to 5 or 6) the animals often respond with a typical and fatal pneu- 
monia; a virus (pneumonia virus of mice) can be isolated (348; 486). 

Another important example is the so-called Theiler's virus (mouse 
encephalitis; 646). This latent virus was detected in course of trans- 
fers for attempted isolation of poliomyelitis virus, and, interestingly 
enough, it causes in mice a syndrome very similar to poliomyelitis in 
man or in anthropoid apes. Theiler's virus and other related viruses 
are widespread in mice. 

The reason for the sudden acquisition of virulence by a latent virus 
upon artificial transfers from one animal to another animal of the same 
species, even within the same colony, is obscure. It is possible that 
if the virus enters an animal very early in the animal's life it can estab- 
lish some sort of equilibrium by which it causes only mild symptoms. 
If large amounts of virus are injected into an adult, the large virus 
inoculum, together with the damage caused by the injection, may 
establish conditions favorable for a sudden increase in virus repro- 
duction leading to destruction of many host cells. 

The problem of latent viruses, however, has broader connotations. 
Instances of latent viruses have often been cited as suggesting a pos- 
sible origin of viruses from normal cell components, which become 
viruses upon being inoculated into other cells. The question of virus 
origin will be discussed more fully in chapter 18. We may mention 
here that virus latency is not a special situation but only an extreme 
one. Any virus or, more generally, any obligate parasite that is very 
destructive to some host must have established in nature a better modus 
vivendi with some other host, in order to insure its own survival. 
Partial latency, then, is almost a necessity for every virus in some host. 
The virus of herpes simplex, for example (114), is present in a very 
large number of individuals, in whom it reveals itself only occasionally 

38 Detection and Identification of Viruses 

by producing "fever blisters" or "cold sores" as a result of some dis- 
turbances of the epidermal cells, mainly around the oral cavity. This 
had been considered an instance of endogenous origin of a virus, until 
it was realized that the virus enters most individuals of a human com- 
munity in early childhood, causing a mild oral inflammation called 
aphthous stomatitis (180). After the child recovers, the virus ap- 
parently remains localized in the neighborhood of the initial point of 
entry and in a state of equilibrium with the host cells, an equilibrium 
which may be broken by external influences. 

A beautifully analyzed case (651), which shows various possible 
degrees of latency, is that of the lymphocytic choriomeningitis virus. 
This virus causes a sporadic form of meningitis in man and also pro- 
duces a meningitic syndrome in apes. The virus was found present 
in a colony of laboratory mice. Upon successive transfer within the 
mouse colony the infection became widespread and nonsymptomatic. 
After a longer period of permanence in the colony the virus infected 
practically every animal, at first by entry in very early age, later even 
by intrauterine transmission. The virus passes through the placenta 
and may even be propagated through the egg. In these mouse 
colonies, then, it is difficult to distinguish the virus from a normal 
cell component, although we know when the virus first entered the 
colony. We also know that the virus is present in endemic form in 
the grey mouse, from which it is occasionally transmitted to the labora- 
tory mouse (31). 



Titration of Viruses 


Just as the presence of a virus is detected through the manifestations 
it produces in a sensitive host, so also the amount of virus activity is 
determined from the quantitative relations between the amount of the 
virus-containing material and the degree or frequency of production 
of specific responses in a host. The purposes of measurements of virus 
activity are manifold. The need for determining the amount of virus 
arises not only in comparing different virus suspensions but also in 
the course of isolation, concentration, and purification of viruses and 
in the study of treatments that may affect virus activity. Moreover, 
methods for measuring virus activity are necessary in any study of virus 

The determination of virus activity presents a number of peculiarities 
that distinguish it from other types of activity measurements, such as 
measurements of chemicals by titration. In chemical titration, what is 
determined is the amount of a given reagent with which an unknown 
amount of a chemical can react. In virus titration the conditions are 
somewhat different, because the virus reproduces in the sensitive host. 
Reproduction results in an amplification of the effects of very small 
amounts of virus, so that often the manifestations produced by small 
or large amounts of virus are indistinguishable. We have, therefore, 
an "all-or-none" effect. A comparison can be made with the growth 
of a bacterial culture, where under favorable circumstances the final 
result visible growth depends only on the presence or absence of 
viable bacteria in the inoculum, and not on their number. 

In virus titration we determine the smallest amount of a virus sus- 
pension capable of producing a suitable manifestation in a susceptible 
host. This manifestation may be a generalized disease or a localized 
lesion, according to the virus investigated and to the purpose of the 


40 Titration of Viruses 

titration. The smallest amount of virus capable of producing a re- 
action is called an infectious unit, and the titer of the original virus 
suspension is given in terms of the number of infectious units per unit 
volume. For example, if the smallest amount of a suspension of in- 
fluenza virus that causes pneumonia in mice upon nasal instillation is 
0.1 ml of a dilution 1/10 6 of an infected lung emulsion, 1 we calculate 
an approximate titer for the emulsion as 10 7 infectious units per ml: 
1/(0.1 X 10- 6 ) = 10 7 . 

In general, titration procedures depend on methods by which inocu- 
lations can be scored as giving a positive or a negative result, as, for 
example, the presence or absence of lysis in a bacterial culture inocu- 
lated with bacteriophage, the presence or absence of an inflamma- 
tory reaction at the site of intradermal inoculation of vaccinia virus 
in rabbit, the presence or absence of paralytic symptoms in a mouse 
injected intracerebrally with encephalitis virus. It is sometimes pos- 
sible to score the presence or absence of virus reproduction inde- 
pendently of the observation of host reactions. In the titration of 
influenza virus by inoculation in the allantoic fluid of the chick em- 
bryo, some fluid from each egg is extracted several hours or days after 
inoculation and is tested for presence or absence of specific hemag- 
glutinating ability (see chapter 13). 

Finally, there are., cases in which the amount of activity can be 
measured from the number of specific responses appearing on a given 
surface as a result of the inoculation of a given amount of virus, as in 
the count of plaques for bacteriophage, in the same way as bacterial 
titers are estimated by colony counts. 

1 The reader should be familiar with the use of decimal notation in dealing with very 
large or very small numbers. ' For example, 1,000,000,000 = 10 9 ; 1/10,000 = 10~ 4 ; 
7,030,000 = 7.03 X 10 6 ; 1/250,000 = 10/2,500,000 = 4/1,000,000 = 4 X 10~ 6 . The ex- 
ponent of 10 in the above expressions is the logic of the power of 10 next below. The 
number 4 X 10" 6 could also be expressed as 10 (log 4) " 6 = 10 ~ 5 398 . In a series of dilutions 
by a factor 2, the decimal logarithms of the concentrations will differ by a constant amount 
log 2, that is, about 0.3 (for example, ^ 10 ~ 3 ; V 10 ~ 06 ) In some cases, the 
reader will encounter the natural logarithmic notation (e n \ e~ n , etc.). The symbol e, 
the base of natural logarithms, equals 2.71828 ; log x 2.3 logio *; logio x 0.43 
loga x. The reader should be thoroughly conversant with the elementary theorems on 
powers, roots, and logarithms (for example, 10* X 10 6 = 10 a+6 ; 10~ = l/(10 a ); 10 0/6 = 
\/10; log a 6 = log c b X log a c). 

CH. 3 

Titration Procedures 



Bacteriophage titration. The simplest example is that of bacterio- 
phages. These are generally titrated by preparing several dilutions 
and plating one or more measured samples of each dilution on nutrient 
medium plates with an excess of sensitive bacteria. Phage and bac- 
teria may be introduced in a surface layer of nutrient agar (269; 316). 
After incubation, the continuous bacterial layer is interrupted by the 


- 320 

J 160 


- 80 


I I I I I I 

2 4 8 16 32 64 128 256 
Relative concentration of phage 

Figure 8. The proportionality between plaque count and concentration of bac- 
teriophage. From: Ellis and Delbriick (205). 

clear areas of lysis produced by phage. The plaque counts are analo- 
gous to colony counts for bacteria, and their values are directly pro- 
portional to the amount of phage plated (figure 8). The titer is 
calculated from the plaque count, the volume of phage plated, and 
the dilution. If 0.1 ml of a dilution 10~ r> gives on the average 230 
plaques, the titer will be 230/(0.1 X 10~ 5 ) = 2.3 X 10 8 units per ml. 
In general, if y plaques are produced by a volume v (in milliliters) of 
a dilution x, the titer n of infectious units per ml is n y/vx. 

The precision of each determination on an individual plate can be 
estimated by remembering that the standard deviation, due to sampling 
errors, is equal to the square root of the average number of plaques 
per plate. The precision, as measured from the ratio ^/n/n between 
standard deviation and count, will be greater the higher the average 

42 Titration of Viruses 

number n of plaques per plate. A limitation is imposed, of course, by 
the fact that the plaques, if too numerous, become confluent and diffi- 
cult to count. 

In estimating the precision of a virus titration involving dilutions it 
must be remembered that, apart from sampling errors resulting from 
random distribution of particles, we have technical errors in the volume 
of the samples transferred in the dilution procedure. If, for example, 
a titration involves 5 dilution steps, each with a probable error of 
9%, the probable error of the whole dilution series will be (9 X V5 )%. 

Several technical points must be observed in carrying out any virus 
titration by the dilution methods. In the first place, a different pipette 
should be used for each successive dilution. This requirement is more 
strict for virus titrations than for chemical or serological procedures, 
because virus activity may be conditioned by the presence of single 
particles, and a pipette that has contained millions of virus particles 
cannot easily be freed of all of them by repeated rinsings. This pre- 
caution seems obvious, but has so often been neglected by beginners 
and others that its neglect has been christened (186) with the special 
name of Pipettenfehler (pipetting mistake). Other important pre- 
cautions involve the choice of a dilution fluid in which virus activity 
does not rapidly deteriorate, and the control of temperature and other 
factors that may affect virus activity. 

Animal viruses. 'Methods analogous to the plaque count for phage 
have been used, particularly the counting of discrete lesions or "pocks" 
produced by several viruses on the chorioallantoic membrane of the 
chick embryo. The maximum number of pocks countable on one 
membrane is about 20-100 for different viruses. Because of the rather 
complex manipulations involved and of the limited applicability of 
this technique, its use has been restricted to a few laboratories (77). 
Even more limited in application is the count of pustules produced on 
scarified areas of rabbit skin (281) or cornea (320), which have been 
used mainly to estimate the potency of vaccines. 

A method recently introduced (196) consists of counting discrete 
lesions or plaques produced on monolayers of specially prepared tissue 
cultures. The "plaque count," exactly analogous to that used with 
phage, is directly proportional to the amount of inoculum. This tech- 
nique should provide an ideal method of titration when it becomes 
applicable to a large variety of viruses. 

The most common titration procedure, however, is the inoculation of 
equal samples of successive dilutions of a virus suspension into sus- 
ceptible animals, followed by the determination of the dilution end 

CH. 3 Titration Procedures 43 

point by scoring each inoculation as positive or negative according to 
the typical host responses. Titers, in infectious units per milliliter, ate 
obtained from the concentration and the volume of the sample inocu- 
lated, assuming the last positive dilution to contain 1 unit. The factor 
between dilutions is usually % or %; the smaller the denominator, 
the more precise is the titer obtained. In titrations of this type, there 
is often an "uncertain zone/' in which a given dilution may give a 
negative result, whereas the next higher dilution gives a positive one. 
Such an "uncertain zone" is generally interpreted as due to the fact 
that an infectious unit consists of one discrete material particle. 2 

The results of a titration by dilution end point can be made more 
accurate by testing several samples of each dilution. For one or more 
dilutions there will be some positive and some negative inoculations, 
and it will be possible to estimate by interpolation the dilution that 
would give 50% positive and 50% negative results. The titers can then 
be expressed in multiples of the "50% infectious dose." This method 
can be applied to a variety of virus titrations (influenza virus in the 
allantoic cavity of chick embryo, neurotropic viruses in the brain of 
mice, etc. ) , whenever the response can be scored as positive or nega- 
tive. The method is illustrated in table 3. 

A refinement was introduced by Reed and Muench (546). The per 
cent of positive responses is calculated not from the actual frequencies 
for any one dilution but from the "accumulated sums," as illustrated 
in table 4. This corresponds to assuming that, if a given individual 
test (for example, a given egg) was positive for a certain dilution, it 
would have been positive if used to test any of the lower dilutions, 
and vice versa for a negative test. This method of calculation is only 
justified, however, if one assumes that the negative responses obtained 
at dilutions at which other inoculations are positive reflect variations 
in host sensitivity. We shall see that there is good reason to believe 
that the negative responses depend partly on the chance absence of 
discrete material particles ( one particle constituting an infectious unit ) 
and do not necessarily reflect variations in host sensitivity. The theo- 
retical justification of the "accumulation method" in virus work appears 
doubtful, although in practice its use has many advantages. 

2 Let us suppose that a dilution 10 - 4 contains 50 infectious units per ml. A 
0.1-ml sample will contain on the average 5 units and very likely will give a posi- 
tive response. A 0.1-ml sample of 10 ~ 5 dilution will contain 0.5 unit on the 
average, and one such sample may or may not contain 1 unit. A 0.1 ml sample 
of 10- dilution will almost always give a negative response, but it may happen 
that the sample of 10 5 dilution tested contains no unit, while 1 unit has been 
transferred to the next dilution, with an anomalous result. 


Titration of Viruses 

Table 3. Virus titration by the 50% end-point method 

Calculation of LD M (50% lethal dose). Data from Parker and Rivers (oil). 







% Positive 

io- 3 

0.1 ml 



io- 4 




10~ 5 





10~ 6 

0.1 ml 




io- 7 

0.1 ml 


50% end-point = 10~ 5 5 . 

0.1 ml of a 10~ 6 5 dilution = 0.1 X 3.16 X 10~ 6 ml = 1 LD 60 . 

Titer in LD 50 per ml = 1/(0.1 X 3.16 X 10~ 6 ) = 3.2 X IO 6 . 

Table 4. Virus titration by the Reed and Muench method 




(sum down) 

(sum up) 

Per Cent 

10~ 3 




io- 4 




10~ 5 






10~ 6 






io- 7 



The 50% end-point can be obtained by graphic interpolation (see -46') . 

Several points must be kept in mind in titrating animal viruses by 
end-point methods. The results are bound to depend not only on the 
amount of virus but also on the route of inoculation, on the host 
animal, and on a number of other less definable factors. We must 
remember that a positive response is a manifestation of successful 
infection. This in turn depends on the proper amount of virus ma- 
terial reaching a susceptible cell, succeeding in infecting it, and 
propagating the infection to other cells. All sorts of host defenses 
(mechanical, humoral, and cellular) tend to reduce the chances for 
successful infection. The titers will then indicate the number of 

CH. 3 Titration Procedures 45 

infectious units under the specific conditions of the test. For example, 
one and the same sample of influenza virus may give a titer of 10 8 
units per ml if titrated in eggs and a titer of 10 units per ml if titrated 
by nasal instillation in mice. This may reflect either the greater re- 
sistance of mice or the presence of two types of virus materials in the 
suspension, with different abilities to attack different hosts. The 
situation is analogous to that of a bacterial suspension giving a higher 
colony count in a favorable medium than in an unfavorable one, where 
many cells fail to proliferate. 

The incubation period method. Partly because of the large numbers 
of host animals needed for titration of certain viruses by the end-point 
method, other titration methods have been devised, none of which is 
of general application. The most successful type is based on the fact 
that for some viruses there is a fairly precise relation between the 
amount of inoculum and the "incubation" time between inoculation 
and appearance of a given symptom. For example, for rabbit papil- 
loma virus Bryan and Beard (103) found that the time interval be- 
tween the inoculation in the rabbit skin and the appearance of 
papillomas is related to the amount of inoculum C x by the relation 
t 2 ti = &(log Ci log C- 2 ). A plot of t versus log C gives a straight 
line, and unknown samples can be compared by reference to a stand- 
ard plot (figure 9). Other workers found and used a similar relation 
between the time of death and the amount of inoculum. Card (248) 
found for certain strains of Theiler's virus (mouse encephalitis) a rela- 
tion of the type (1/fi) - (I/**) = fo(log Ci - log C 2 ) between the 
inoculum size and the time of onset of paralysis. Card pointed out 
that a relation of the type 1 2 ti = b(\og C t log C 2 ) can be ration- 
alized if we postulate (a) that symptoms (or death) appear when the 
virus concentration in the host has reached a constant level; and 
(&) that virus reproduction is an approximately exponential process, 
similar in its overall course to the reproduction of bacterial cells in a 
growing culture. 3 The relation (1/fi) - (l/t 2 ) = fc(log C 2 - log Ci) 

5 Let us call Y the critical number of virus units present when symptoms appear, and 
let us postulate a relation between virus amount y and time t such that 

y - *' In y = kt 

Let T be the time needed to reach the amount Y of virus from an inoculum containing 
1 unit. Then, for an inoculum containing C\ units, the time t\ needed to reach the amount 
Y will be shorter than T by the time that would be needed to go from 1 to Ci units: 

In Y - In Ci = k[T - (T - fc)) - bi (a) 

Similarly, for an inoculum fg: 


Titration of Viruses 

could reflect an accelerated logarithmic reproduction, for example, a 
progressively more successful overcoming of host defenses as the virus 
concentration increases. No direct confirmation for such a rate of 
virus reproduction in animals has yet been offered. 

Plant viruses. Titrations are generally done by the count of necrotic 
or other lesions, a method introduced by Holmes in 1929 for tobacco 







I I I I I I I I I I I I 71 I I 

I I I I I 

J I I i 1 I I I I I 


p co 

00 00 


Figure 9. The inverse proportionality between the incubation period and the 
logarithm of the amount of papilloma virus inoculated, pi decimal logarithm 
of the dilution factor (for example, pi 7.4 corresponds to a dilution 1:10 74 ). 
From: Bryan and Beard (103). 

mosaic virus (333) and extended later to many other viruses. Serial 
dilutions of a virus suspension are rubbed onto the surfaces of the 
leaves of susceptible plants. The leaves are held flat against the hand 
of the experimenter and rubbed with a piece of cheesecloth or some 
other suitable applicator previously immersed in the virus suspension. 
The number of lesions produced by a given dilution gives a measure 
of the amount of virus present (figure 10). The method requires a 
great deal of standardization and practice before meaningful results 

In Y - In (\ = k[T - (T - fe)J = &2 (W 

Subtracting a from 6 we obtain : 

In Ci - In C 2 = k(h - t\) 

CH. 3 

Titration Procedures 


can be obtained. Even the amount of pressure exerted in rubbing 
the leaf has an effect on the count of lesions, since only virus particles 
that penetrate the leaf cells through local abrasions can proliferate. 
Abrasive powders such as carborundum are therefore often mixed with 
the virus suspension or dusted on the leaves previous to inoculations. 
A number of improvements have been introduced. For example, 
the half -leaf method consists in comparing the number of lesions pro- 
duced by two samples rubbed on opposite halves of the same leaf or 


12 16 20 
Units of antigen 

Figure 10. The relation between the count of necrotic lesions produced by 
tobacco mosaic virus on Nicotiana glutinosa leaves and the concentration of virus 
in the inoculum. The virus concentration is given in units of virus antigen as 
determined by precipitin tests. From: Beale, Contrib. Boyce Thompson Institute 
6:407, 1934. 

of a series of leaves. It was found that results on opposite halves of 
the same leaf are fairly consistent, whereas wide variations occur from 
plant to plant and from leaf to leaf of the same plant, depending in 
part on the position of the leaf on the stem. The Latin square method 
(692) compares a series of samples by rubbing them on different leaves 
of different plants in such a way that each sample is tested in various 
positions on different plants. More elaborate methods are used when 
only limited numbers of plants or leaves are available (691). It is 
stated that about 50 plants are needed for a titration precise to within 
10-50%. Greater precision can be obtained by the use of special ex- 
perimental designs (618a). 

In general, the number of necrotic lesions is not a linear function of 
the concentration of virus, but follows the dilution curve illustrated 
in figure 10. This is due in part to the fact that the maximum number 
of lesions obtainable on a leaf, using a standard technique, is not 
unlimited, but at most is equal to the number of wounds through which 

48 Titration of Viruses 

the virus can enter. The correct measure of activity, therefore, is the 
ratio of the number of lesions obtained to the maximum number ob- 
tainable with the most concentrated sample. For comparison of dif- 
ferent samples it is desirable to work in the range of concentrations in 
which the curve representing the dependence of the number of lesions 
on the virus concentration has its steepest slope, since in that range the 
method is most sensitive to concentration changes. 


The infectious unit, defined as the smallest volume of a virus sus- 
pension that can elicit a positive response, must contain a certain 
amount of specific virus material. What does this minimum amount 
consist of? A priori, we can conceive of two extreme possibilities. 
On the one hand, a positive response may occur whenever enough virus 
material is inoculated to give a certain minimum concentration. This 
is analogous to the requirement for a minimum concentration of a 
chemical poison to produce toxic symptoms. On the other hand, a 
positive response may be conditioned by the chance that the inoculum 
contains at least one individual virus particle that multiplies and causes 
the specific manifestations. An obvious analogy can be drawn with 
bacterial growth, which is conditioned by the presence and reproduc- 
tion of at least one viable bacterial cell in the inoculum. If the second 
hypothesis is correct, then the titers expressed in infectious units are 
indeed numbers of individual units, rather than multiples of minimum 
concentrations. The titers need not express the full numbers of virus 
particles, since some of the particles may fail to manifest themselves, 
just as some of the bacteria in an inoculum may fail to multiply. 

The decision is easily reached in the case of bacteriophages. As 
already mentioned, the count of bacteriophage plaques is completely 
analogous to the colony count for bacteria. The relation between 
plaque numbers and amounts of phage plated is strictly linear (205; 
307), as shown in figure 8. This can only be interpreted by assuming 
that one plaque results from the action of a single independently dis- 
tributed material particle. If more than one particle had to be present 
in one place to produce a plaque, the plaque numbers should increase 
much faster than linearly with increasing amounts of phage plated 
(431). Similar considerations apply to plaque counts of animal viruses 
in tissue cultures (196). 

Statistical analysis. There is an equation that expresses the dis- 
tribution of independently distributed material particles in samples 

CH. 3 Statistical Interpretation of the Infectious Unit 49 

taken from a suspension of such particles. This is the so-called Poisson 
distribution (524): 

K r e~* 

Pr - HI 

where s is the average number of particles per sample, r the actual 
number in a given sample, r\ is the product r X (r 1) X (r 2)X 
3x2 ("factorial" of r), and p r is the probability of having r 
particles in a given sample (and therefore also the expected fre- 
quency of samples containing r particles). For r = 0, p = e~ 8 (since 
()! 1). This is the frequency of samples without any particles. 
The frequency of samples with 1 particle is p ~ se~*\ of those with 
2, p* = (s 2 /2)e~*, and so on. 

The frequency distribution of plaques in series of plates inoculated 
with equal amounts of several dilutions of a phage suspension has 
actually been determined and found to agree with the distribution 
expected from equation 1 (205), as shown in table 5. This represents 

Table 5. Frequency distribution of bacteriophage plaques 

From Ellis and Delbruck (205) 

Thirty-three 0.1-ml samples of a dilute phage preparation were plated for 
plaque counts. The calculated distribution corresponds to equation 1. 

p r (Experimental) p r (Calculated) 

plaques on 13 plates 



1 plaque on 14 plates 



% plaques on 5 plates 



3 plaques on 1 plate 



4 plaques on plates 



1.002 1.000 

direct evidence that the number of plaques represents the number of 
independently distributed phage particles. 

Equation 1 can also be applied where the response is of the all-or- 
none type, for example, presence or absence of disease in inoculated 
animals. The question we ask is: When a positive response occurs, 
does it depend on the action of a single individual particle, or does it 
result from the presence of a minimum number of particles (greater 
than 1)? 


Titration of Viruses 

Let us write the probabilities of having more than 0, more than 1, 
more than 2 particles per sample: 

P r >o = 1 ~ Po = 1 ~ e~ s 

P r>l = 1 - ( Pl + po) = 1 - (s + l)e 

Pr>2 = 1 - (P2 + Pl + Po) = 1 - [(* 

+ S + l]e 



These expressions also represent the frequencies of samples that con- 
tain at least 1, or 2, or 3 particles. The dependence of the pro- 
portion of samples with more than a given number of particles on the 
relative number of particles is illustrated in the plots of figure 11. 
The plots give the values of Pr>n versus log 5 for the various values 
of n. 



-3-2-1 1 2 

Figure 11. The theoretical dependence of the frequency of positive samples 
( samples with at least n particles ) on the concentration of particles in an inoculum. 
Curves 1, 2, 3, 4 represent the expected frequencies for n = 1 or 2 or 3 or 4. 
Abscissa: logarithm of concentration (empirical scale). Ordinate: frequency of 
positive samples. Note the increasing steepness of the response-concentration 
curve for increasing values of the minimum number of particles required. From: 
Lauffer and Price (404). 

The reason for the shape of the curves in the plot of P versus log s 
is clear: The greater the number of particles required, the steeper will 
be the dependence of the frequency of positive responses on amount of 
virus because of the faster transition from the point where all samples 
contain more than the minimum number of particles to the point where 
none does. 4 

4 Suppose we distribute 1400 marbles at random in a series of 10 boxes, and we 
score as "positive" the boxes that receive at least 100 marbles. Very likely, all 
boxes will receive more than 100 (all "positive"). If we distribute 700 marbles at 
random in 10 boxes, very probably all of them will have fewer than 100 (all 

CH. 3 Statistical Interpretation of the Infectious Unit 51 

To apply these curves to virus titration, let us prepare a series of 
virus dilutions and inoculate for each dilution a series of equal samples 
for infectivity test. The (unknown) average numbers of virus particles 
in each series will be proportional to the concentration. The fre- 
quency of positive inoculations is obtained for each series and is 
plotted versus the logarithm of the dilution. The dependence of this 
frequency of positive responses on the dilution should represent the 
dependence of the presence of the minimum necessary number of 

Figure 12. Experimental frequency of positive responses to inoculation of vari- 
ous amounts of a vaccinia virus preparation in rabbit skin. Notation as in figure 
11. The solid line is curve 1 of figure 11. From: Lauffer and Price (404). 
Data from Parker (509). 

particles on the average number of particles per sample. This can 
be compared with the theoretical curves of figure 11. Such compari- 
sons, from data for various viruses, are shown in figures 12 and 13. 
The experimental data as a whole fit much better the curve correspond- 
ing to the requirement for at least 1 particle than those corresponding 
to the requirement for at least 2 or more particles. 

With bacteri&phage, the experiments consist of placing equal sam- 
ples with various concentrations of phage into tubes of bacterial culture, 
and recording the presence or absence of lysis. The results ( see figure 

"negative"). Thus, dividing the average number of marbles per box by 2 has 
reduced the frequency of "positive" cases from 10/10 to 0/10. 

Suppose now that we consider as "positive" the boxes with at least 1 marble. 
We distribute 14 marbles in 10 boxes (1.4 average per box). The expected 
number of boxes without marbles is 2 ( = 10 X e~ lA ). If we distribute 7 marbles 
only, the expected number of empty boxes becomes 4. Cutting the average 
number of marbles per box in half has reduced the number of "positive" boxes from 
8 to 6. Thus, the smaller the number required for a "positive" score ( 1 instead 
of 100), the more gradual becomes the transition from the input that makes all 
scores "positive" to the one that makes them all "negative." 

52 Titration of Viruses 

13) are in excellent agreement with the frequency distribution expected 
from equation 2 (r 0; 1 particle sufficient to produce lysis) (213; 
431). This means that 1 phage particle is sufficient to initiate mass 
lysis in a susceptible bacterial culture, or, in other words, that the 
infectious unit consists of 1 material particle. 

We must realize, however, that statistical tests of this nature do not 
tell us that every virus particle will succeed in starting infection. It 



0.1 1 

Relative amounts of bacteriophage 

figure 13. The frequency of positive response (lysis) in sets of bacterial cul- 
tures inoculated with various amounts of bacteriophage. Each point corresponds 
to a set of 160 cultures. The solid line is equivalent to curve 1 of figure 11. 
Modified from: Feemster and Wells (213). 

is clear that, even though 1 particle can cause infection, not every 
particle will succeed in doing so, either because of intrinsic differences 
in the particles themselves or because of differences in the conditions 
of host-virus interaction, such as, for example, the presence of in- 
hibitors that may inactivate some of the particles of the inoculum. The 
statistical analysis tells us only that infection can be initiated by a 
single particle and that it does not require the collaboration among 
several particles. 

For bacteriophage, the plaque count titers can be compared with the 
actual numbers of characteristic phage particles recognizable in the 
electron microscope (see chapter 4). The agreement is very good, 
the number of actual particles being between 1 and 2.5 times the 

CH. 3 Statistical Interpretation of the Infectious Unit 53 

plaque counts. This means that at least 40% of the phage particles 
recognizable in the electron microscope, and often 100$ of them, suc- 
ceed in forming plaques (449). 

Experiments with animal viruses (509; 510; 512) have given similar 
results. In practically all cases the results fit the curve corresponding 
to equation 2(1 particle to infect). With a very virulent strain of 
vaccinia virus titrated in the skin of rabbits, Parker (509) found an 
excellent agreement with the hypothesis that the infectious unit con- 
sists of 1 particle (figure 12). Often, however, for animal viruses the 
frequency of positive results increases even more gradually with in- 
creasing amount of virus suspension than expected from equation 2. 
Deviations of this type can be explained by assuming that, although 
1 particle is sufficient to start infection, other factors increase the rela- 
tive chance of positive response from small inocula. For example, 
reversible aggregation at higher concentrations, reversible combination 
with inhibitors in the inoculum, or variations in host resistance could 
produce the observed results (see 248). 

For vaccinia virus the conclusion that a single particle can initiate 
infection is supported in some cases by the close correspondence 
between infectious titers and numbers of characteristic elementary 
bodies in the same suspensions (about four times as many bodies as 
infectious units; 606). For viruses such as that of rabbit papilloma 
tested in rabbit's skin, however, the discrepancy is much greater, up 
to 10 7 or 10 8 particles per infectious unit. This in itself does not 
exclude the hypothesis that 1 particle is an adequate infectious unit; 
the other 10 7 or 10 may either be inactive or may have very little 
chance of successfully infecting a cell. Nevertheless, Bryan and Beard 
(104) have suggested that in papilloma the response frequency curve 
does not correspond to the frequency with which at least 1 virus 
particle is successful in producing infection, but is an expression of 
the distribution of sensitivity of different cells or groups of cells to 
the virus. They postulated a distribution of sensitivities similar to 
that found for response to many drugs (a normal distribution of the 
logarithms of the sensitivity levels, expressed as the minimum effective 
drug concentration; 247). On this basis an expression can be derived 
which fits virus titrations as well or better than equation 2. These 
authors have, therefore, proposed the general conclusion that varia- 
tions in host sensitivity to viruses are alone responsible for the shape 
of the frequency distribution of positive responses as a function of the 
amount of virus inoculated. This conclusion is not applicable to phage 
titration and probably not even to vaccinia virus. The good cor- 

54 Titration of Viruses 

respondence of particle numbers and infectious unit titers (449; 606) 
indicates that, with inocula that give some positive and some negative 
inoculations, the presence or absence of particles must play the major 
role. For other animal viruses, the sensitivity distribution of the hosts 
may mask or distort the effects of the statistical distribution of the 
individual virus particles. 

Similar considerations have been applied to the titration of plant 
viruses by necrotic lesion count. Several authors (35; 404; 693) have 
shown that the ratio y/N of the number y of lesions obtained with a 
given concentration to the maximum obtainable number N of lesions 
obeys the relation y/N = 1 e~*. This equation is analogous to 
equation 2 and indicates that infection results from the action of a 
single particle at one point of entry. Some frequently occurring de- 
viations have been explained (35) as due to reversible aggregation of 
virus particles at the highest concentrations. The ratio (infectious 
units)/ (virus particles) is generally very low, between 1/1000 and 
1/10 7 . This is not surprising, in view of the rather inefficient method 
of inoculation by leaf rubbing. Kleczkowski (375) has published 
results on tobacco mosaic virus similar to those of Bryan and Beard 
(104) on rabbit papilloma. The suggestion was made that the titration 
curve of plant viruses may reflect in part or exclusively the distribution 
of sensitivity of the sites of virus entry to various virus concentrations, 
rather than the presence or absence of individual successful virus 

The question is at present unsettled. As a whole, it seems likely 
that for most viruses 1 particle is sufficient to initiate infection. For 
some viruses the titration results may not be a measure of the num- 
ber of successful individual particles, but simply that of the virus 

Even in the most favorable cases, the titer of a virus suspension, 
expressed in infectious units per unit volume, corresponds to the 
number of particles that are successful in producing infection under 
the particular conditions employed in the test. The numerical rela- 
tion between successful particles and actual virus particles cannot be 
decided by titration or by any other statistical tests, but only by actual 
counts or determinations of particle number. These counts, in turn, 
cannot distinguish between active and inactive particles, nor between 
particles differentially infectious for different hosts or for different 
routes of inoculation. At best, titration results give numbers of statisti- 
cal particles, whose relation to the number of actual particles involves 
unknown factors of proportionality. 


Size and Morphology 
of Virus Particles 

We have seen that virus preparations contain material particles with 
specific activity. The methods for the demonstration, identification, 
and physical study of these specific particles represent an important 
phase of virus research, involving the use of physical and chemical 
procedures which are for the most part derived from colloid chemistry 
and protein chemistry. 

Preparation of crude virus suspensions. If the native virus material 
is already a fluid, for example, a phage lysate or a virus-containing 
body fluid, the only requirement is the elimination of bacterial con- 
taminants by filtration through any bacteriological filter that does not 
retain the virus. The choice of filter is mainly a matter of experience; 
diatomaceous earth or sintered glass filters are generally suitable. If 
the virus is mostly present in intracellular form in a tissue, the tissue 
is minced and ground with sterile sand or ground glass or otherwise 
homogenized to release the virus into the surrounding fluid. The pulp 
is suspended in a suitable medium. Most viruses are quite stable in 
media of composition similar to that of host tissue, which, for animal 
viruses, is approximated by bacterial culture media such as meat- 
infusion broth. Plant viruses are generally isolated by breaking up 
the plant cells so that the virus is released with the plant juice. This 
juice should be brought near neutrality by addition of proper buffers. 

Crude virus suspensions are generally stored either in a refrigerator, 
or by quick-freezing and storing at temperatures as low as 70 C 
(344), or by drying in vacuo from the frozen state. Quick-freezing 
avoids the formation of ice crystals and minimizes protein denatura- 
tion. The fact that all known viruses are rapidly inactivated at tem- 
peratures below 100 C makes it possible to destroy the infectivity of 
viruses by the standard sterilization procedures of bacteriology. 


56 Size and Morphology of Virus Particles 

Identification of virus particles. Once a suspension containing virus 
activity is available, we proceed to search it for some material to which 
the specific activity can be attributed. Ideally, such material should 
be present in infectious preparations and absent from noninfectious 
control preparations. Its amount should parallel the amount of virus 
activity, and its alteration should be accompanied by loss of activity. 
By a variety of methods, it has been found that virus activity is gen- 
erally associated with particles characteristic for each individual virus, 
whose sizes vary all the way from less than the size of small bacteria 
(0.3 /A) to the size of medium-size protein molecules ( 10 m/x). Table 6 
summarizes data for a number of viruses. 

It will be useful to recall briefly some of the units of measure 
encountered in virus work. A micron (/*) is 10 ~ 4 cm; a millimicron 
(m/x), 10 ~ 7 cm; an angstrom (A), 10 ~ 8 cm. The angstrom unit is 
generally employed in measurements of atomic and molecular dimen- 
sions. For example, the distance between the centers of two carbon 
atoms in the ethane molecule is about 1.5 A; the distance between 
amino acids in the polypeptide chain of silk protein is 3.5 A. For 
large molecules or colloidal particles, we use the millimicron unit. 
The spherical molecule of egg albumin, containing about 350 amino 
acid residues, has a radius of 2.7 m/* (27 A). For a spherical particle 
of radius r, the voluine is equal to %7rr 3 . If the volume and density 
are known, we can calculate the mass of a particle. The mass can be 
expressed either in standard mass units, for example, the gram and its 
submultiples, or in multiples of the unit of molecular weight, which is 
taken as % 6 of the mass of the oxygen atom, that is, 1.65 X 10 ~ 24 
gram. Expressed in this way, the mass gives the molecular weight 
of a molecule having the same mass as our particle. A convenient unit, 
corresponding to 10 6 units of molecular weight, has been given the 
symbol Mh (521). As an example, a spherical virus particle, 20 m/t 
in diameter, consisting of material with density 1.1 grams per cm 3 , has 
a volume of (4 X 3.14 X 10 3 )/3 m^ 3 = 4190 m/* 3 = 4.19 X 10~ 18 cm 3 , 
and a mass of 4.19 X 10~ 18 X 1.1 gram = (4.19 X 1.1 X 10- 18 )/ 
(1.65xlO- 24 Xl0 6 ) Mh = 2.8 Mh. The inverse of the unit of 
molecular weight is N = 6.06 X 10 23 , the Avogadro number. raN/10 
gives the mass in Mh units, if m is the mass in grams. 


The resolving power of the ordinary light microscope, defined as 
the inverse of the smallest distance at which two objects can be seen 

Table 6. The sizes of the particles of representative viruses 



Linear Dimensions, millimicrons 

Microscopy * 

Diffusion f 

tion J 

Animal viruses 

Influenza (A, B, swine) 

Spheres (and fila- 




Fowl pox 


320 X 260 

Molluscum contagiosum 


330 X 260 

Chicken pox 

300 X 230 

Herpes zoster 


330 X 260 

Vaccinia, variola 


260 X 210 

Rabbit myxoma 


290 X 230 

Psittacosis, lymphogranuloma 



venereum, feline pneumoni- 





Rabbit fibroma 




Rabbit papilloma (Shope) 




Rous sarcoma 



Milk factor (mammary carci- 



noma of mice) 

Louping ill 







Mouse encephalitis 



Foot-and-mouth disease 




Equine encephalomyelitis 



Japanese (type B) and 



St. Louis encephalitis 

Pneumonia virus of mice 



Yellow fever 




Polyhedrosis of silkworm 


280 X 40 

Granulosis (capsule disease) of 


270 X 40 


Plant viruses 

Tobacco mosaic 


300 X 15 

Potato virus X 


16 (thickness) 

Potato yellow dwarf 

Spheres (?) 


Tomato bushy stunt 




Southern bean mosaic 




Tobacco necrosis, Rotham- 



stead strain 

Turnip yellow mosaic 




Coli-phage Tl 

Spermlike, prismatic 

Head: 50 


Tail: 150 X 10 

Coli-phage T2, T4, T6 

Spermlike, prismatic 

Head: 95 X 65 


Tail: 100 X 25 

Coli-phage T5 

Spermlike, prismatic 

Head: 65 


Tail: 170 X 10 

Coli-phage T3, T7 

Prismatic (short tail) 


Coli-phage 813 


Coli-phage Kottmann 


100 X 35 

* Electron microscopy gives values for dry and possibly flattened particles, 
t Sedimentation and diffusion methods give values for hydrated particles. 
$ Ultrafiltration data are approximate within a factor 1.5-2. 


58 Size and Morphology of Virus Particles 

as separate and distinct, is given by: R.P. = 2(N.A.)/A, in which A is 
the wavelength of the light and N.A. is the numerical aperture of the 
objective lens. For visible light, A = 400-800 m/x. Since the best 
numerical aperture obtainable is approximately 1.4, the limit of reso- 
lution is of the order of 200 m/i. This is just below the limiting porosity 
of ordinary bacterial filters that allow viruses to pass, and only the 
largest viruses can be visualized with the ordinary light microscope. 
The use of staining techniques improves the visibility of large virus 
particles both by improving contrast and by increasing the actual size 
of the particles through deposition of staining material. 

In 1887, Buist (106) observed particles near the limit of visibility in 
smears from the contents of pustules of vaccinia and smallpox viruses 
after staining with the basic dye carbol fuchsin. These particles were 
later studied by Paschen (514), using either carbol fuchsin or Giemsa 
staining solution ( methylene azur and methylene azur eosinate ) . The 
work of Paschen established the specific relation between virus ac- 
tivity and the presence of these bodies, which he called elementary 
bodies. This term has remained in general use to indicate particles 
associated with activity of certain viruses barely visible through the 
light microscope (figure 14). The elementary bodies of the vaccinia- 
smallpox virus group are often called Paschen bodies. 

Generally, all techniques for staining elementary bodies employ 
either strongly basic dyes or metallic silver precipitated from a col- 
loidal solution of silver hydroxide onto particles previously treated 
with a mordant (Morosow's method; see 655). Elementary bodies 
in purified suspensions may actually be counted in stained prepara- 
tions. The identification of elementary bodies in stained preparations 
hinges on the possibility of associating their presence with the presence 
of virus activity and may be complicated by the presence of foreign 
particles of similar size. 

Dark-field microscopy is based on the visualization of particles by 
the light they scatter into the microscope when illuminated with an 
oblique beam of light which would not enter the objective. It permits 
visualization of particles as small as 50-100 m/x without revealing any 
details of size or shape. It has been found useful in the study of the 
evolution of animal virus infection within infected cells in the living 
state (figures 77, 78) (89). 

The resolving power of the microscope can be improved by using 
ultraviolet light instead of visible light. The resolving power increases 
as the wavelength of the light diminishes. The main difficulty is the 
requirement for expensive quartz lenses or for mirror optical systems 

Figure 14 (left). Particles of vaccinia virus (Paschen bodies) stained by Moro- 
sow's method. Courtesy Dr. G. J. Buddingh, Louisiana State University, New 

Figure 16 (right). Molluscum contagiosum virus, shadowed preparation. From: 
Rake and Blank, /. Invest. Dermatol 15:81, 1950. Courtesy Dr. G. W. Rake, 
Squibb Institute for Medical Research, New York. 

Figure 15. Vaccinia virus, unshadowed preparation. From: Green, Anderson, 
and Smadel (274). Courtesy Dr. T. F. Anderson, University of Pennsylvania, 

60 Size and Morphology of Virus Particles 

since ordinary glass is opaque to ultraviolet light, Moreover, the eye 
being insensitive to ultraviolet, the image must either be observed on 
a fluorescent screen or photographed. Ultraviolet microscopy was ex- 
tensively used by Barnard (38) to observe virus particles. Recently, 
it has been employed by Swedish workers as a microchemical tool to 
measure the specific absorption of virus materials for certain wave- 
lengths and to study quantitatively the changes in this material within 
infected cells (357). 

Hagemann (286) introduced into virology a modification of ultra- 
violet microscopy, fluorescence microscopy. The material to be studied 
is stained with some dye (primulin, thioflavin) which, when illumi- 
nated with ultraviolet light, gives out a greenish-yellow fluorescent 
light. The requirement for quartz optics is reduced to that for a con- 
denser and a microscope slide. Virus elementary bodies, in particular 
those of vaccinia and psittacosis, can be visualized and counted, but 
the method gives no useful information as to their size or shape. The 
recently introduced phase contrast microscope (552) has not yet 
yielded results of particular significance in the study of virus particles. 

Electron microscopy. The electron microscope (see 153; 688) is 
based on the principle that a beam of electrons can be considered as 
a wave system possessing a characteristic wavelength, which depends 
on the speed of the electrons. The faster the electrons, the shorter 
the wavelength. In the microscope, electrons are emitted by a hot 
filament, accelerated by a potential difference of 50 to 150 kv, and 
reach a uniform velocity at which their associated wavelength is much 
shorter than that of light. The electron beam is focused by electro- 
static or, more commonly, by magnetic fields, much as a light beam 
is focused by lenses. The beam is sent through a specimen, and an 
enlarged electronic image of the specimen is projected onto a fluores- 
cent screen or a photographic plate. All electrons that have passed 
through an area of the specimen will be focused onto a corresponding 
enlarged area of the image, except those that have been deflected and 
scattered away by interaction with the electrons contained in the atoms 
of the material under investigation. The probability that an electron, 
in passing through a unit thickness, is deflected, is proportional to the 
number of electrons per unit thickness. Therefore, the loss of electrons 
in a given region of a specimen will depend on the thickness and 
density of the region and on its chemical composition. Particularly 
important is the presence of heavy atoms, which contain more electrons 
per unit volume than lighter atoms. Among the elements present in 

CH. 4 Visualization of Virus Particles 61 

substantial amounts in biological materials, phosphorus and sulfur 
are the most absorbent for electrons. 

Since all substances, including air, can scatter electrons, the bio- 
logical specimens must be examined in a high-vacuum chamber con- 
taining the electron source, the specimen, and the photographic plate. 
This places a severe limitation on the use of the electron microscope, 
since the object must be completely dry. No observation of material 
in the living state can be made. The specimen is supported by a thin 
membrane of cellulose nitrate (collodion) or other plastic material, 
whose thickness (less than 200 A) and lack of heavy atoms make it 
fairly transparent to electrons. Thin films of metals of low atomic 
number such as beryllium have also been employed. The average 
absorption of biological material for electrons is such that hardly any 
detail can be distinguished when an object is thicker than about 0.5 

tO 1 fji. 

Further information can be obtained by the "shadowing" device 
(679). A metal is evaporated in a vacuum chamber containing the 
specimen, in such a way that metallic atoms hit the specimen at a 
small angle. The shadow effect results from the fact that the object 
prevents the metal from reaching the surface of the specimen in an 
area, which appears on a negative as a shadow of the object. The 
length and shape of the shadow reveal the height and surface irregu- 
larities of the object. 

Another method is the use of replicas, that is, of plastic films formed 
by pouring a solution of plastic material on the surface of the object. 
The film formed by evaporation of the solvent is then stripped off the 
surface, and the pattern of ridges and valleys, corresponding to the 
depressions and reliefs on the surface of the object, can be studied 
with the electron microscope. This method has been particularly 
useful in the study of details of crystal surfaces. Pseudoreplicas, in 
which the stripped film carries with it a layer of the material under 
study, are often more useful in virus studies. 

The electron microscope has made it possible to observe the detailed 
morphology of virus particles (figures 15-26). The Paschen bodies of 
vaccinia (figure 15) have a brick-shaped appearance with areas of 
greater opacity, corresponding either to areas of greater condensation 
or to less collapsible regions (274). The darker, more absorbent areas 
might also represent greater concentrations of nucleic acid, and, there- 
fore, of phosphorus atoms. Treatment with alkali, which removes large 
amounts of nucleoproteins from the elementary bodies, leaves a less 
absorbent "ghost," possibly an empty membrane. 


Size and Morphology of Virus Particles 

Figure 17. Feline pneumonitis virus, shadowed. From: Hamre, Rake, and 
Rake, /. Exp. Med. 86:1, 1947. Courtesy Dr. G. W. Rake, Squibb Institute for 
Medical Research, New York. 

Figure 18. Spontaneous mouse encephalitis virus, strain FA, shadowed. From: 
Leyon, Exp. Cell Res. 2:207, 1951. Courtesy Dr. H. Leyon, University of Uppsala, 

CH. 4 

Visualization of Virus Particles 


Figure 19. Influenza A, strain PR8. Virus filaments and particles in tissue 
cultures. From: Murphy, Kurzon, and Bang, /'roc. Soc. Exp. Biol. Med. 73:596, 
1950. Courtesy Dr. F. B. Bang, Johns Hopkins University, Baltimore. 

Figure 20. Tobacco mosaic virus. 
Courtesy Drs. C. A. Knight and R. C. 
Williams, Virus Laboratory, University 
of California, Berkeley. 

Figure 21, Potato virus X. Cour- 
tesy Dr. E. van Slogteren, Laboratorium 
voor Bloemollenderzoek, Lisse, Holland. 


Size and Morphology of Virus Particles 

Figure 22. Tomato bushy stunt virus, air dried. Note the tendency of the parti- 
cles to orient in close-packed array under the influence of surface tension. Cour- 
tesy Dr. R. C. Williams, Virus Laboratory, University of California, Berkeley. 

The particles of each virus are often almost perfectly homogeneous 
in size. This homogeneity cannot be attributed simply to a selection 
during the preparation of the suspensions, since the methods most 
employed for purification, such as differential centrifugation, are not 
very selective for size. The elementary bodies have a narrower size 
distribution than that of bacterial cells from a growing culture, a fact 
that had suggested quite early that they may reproduce by some 
method other than increase in size followed by fission. 

A comparison of the morphology of the various virus particles illus- 
trates the basis for the use of morphological criteria in the classification 
of viruses ( 568 ) . Viruses of the smallpox-vaccinia group have similar 
particles, indistinguishable from those of mouse ectromelia, which, on 
the basis of biological considerations, has recently been shown to be 
related to the vaccinia group (215). The viruses of the influenza 
group have indistinguishable particles. The same is true of those of 
the psittacosis-lymphogranuloma group. 

CH. 4 Visualization of Virus Particles 65 

Figure 23. Phage Tl, air dried. Courtesy Dr. R. C. Williams, Virus Labora- 
tory, University of California, Berkeley. 

The particles of several plant viruses are shown in figures 20-22. 
For various bacteriophages, the morphology of the specific particles 
(figures 23-26) is closely correlated with criteria of genetic and sero- 
logical relationship. The tadpole-like particles of many phages are 
among the most beautiful objects revealed by the electron microscope. 
Their complex morphology is an indication of their elaborate organi- 
zation. Transparent ghosts with intact tails can be obtained from 
tailed phages by a number of treatments, including osmotic shock (20; 
figure 27). This suggests the removal of a portion of the virus ma- 
terials upon lowering of the salt concentration. The "ghosts" are now 
known to contain most of the protein and none of the nucleic acid of 
thephage (313). 

Bacteriophages illustrate the possibility of identifying and char- 
acterizing virus particles in the electron microscope through some 
specific manifestation of their interaction with host cells. Suspensions 
of different bacteriophages, derived from the same host cells, show 
particles characteristic for each phage. By taking "action shots" of 


Size and Morphology of Virus Particles 

Figure 24. Phago T4, frozen-dried. Notice the geometrical features of the 
heads. Courtesy Dr. R. C. Williams, Virus Laboratory, University of California, 

Figure 25. A mixture of phages T6 and T3, frozen-dried. Note the short ap- 
pendage on the bipyramidal particles of T3. Courtesy Drs. D. Fraser and R. C. 
Williams, Virus Laboratory, University of California, Berkeley, 

CH. 4 Visualization of Virus Particles 67 

Figure 26. Phagc T4, unshadowed. Courtesy Dr. T. F. Anderson, University 

of Pennsylvania, Philadelphia. 

preparations dried at intervals after mixing phage ancj bacteria, it can 
be seen that the characteristic particles become adsorbed onto bac- 
teria (440; figure 29). Upon bacterial lysis, characteristic particles 
are liberated in amounts corresponding to the known rise in phage 
activity (448) (figures 30, 31). 

Sometimes, virus particles can be identified by the demonstration of 
some specific reaction, for example, adsorption of the viruses of the 
influenza group on the surface of specifically agglutinable red cells 
(figure 32). The reaction of virus particles with specific antibody has 
been demonstrated in the electron microscope, either as a specific con- 
glutination of the characteristic particles or as an increase in thickness 
of the virus particles upon coating by antibody molecules (22)1 The 
thickness of the rod-shaped tobacco mosaic virus increases upon treat- 
ment with antibody by an amount supposedly corresponding to twice 
the length of an antibody molecule. This suggests a radial orientation 
of the antibody molecules perpendicular to the axis of the virus particle 

The electron microscope can be used to count the number of viru'a 
particles in a suspension (677). The virus suspension is mixed with 
a suspension of particles of polystyrene latex in known concentration. 


Size and Morphology of Virus Particles 

FtgureTTTTeltlTTIiage 'T5, unsKaHov^d, after Y m > o *|j^^5 c Y^ e ^^ e ^~ ^ u _ 
tice the normal skin and tail and the empty head. Courtesy Dr. T. F. Anderson, 
University of Pennsylvania, Philadelphia. 

Figure 28 (right). Immature particles of phage T2 ("doughnuts") produced 
under the action of proflavine. Note the resemblance with empty phage heads 
(see figure 27). Courtesy Dr. T. F. Anderson, University of Pennsylvania, Phila- 

The latex particles are spherical, extremely homogeneous in size, and 
easily countable. The mixed suspension is sprayed in a fine mist with 
a nebulizer, and droplets are collected on the specimen holder, are 
allowed to dry, and are examined. Whenever a droplet fits within 
one electron microscopic field, all the particles of either latex or virus 
it contains are counted ( figure 33 ) . From their relative numbers, we 
can then calculate the concentration of virus particles in the suspension. 
The estimate can be made quite accurate by increasing the numbers 
of particles counted. This method was used, for example, to compare 
the number of phage particles with that of the infectious units of 
phage (449) and to calculate the weight of the particle of tobacco 
mosaic virus. This was done by counting particles in accurately 
weighed preparations of pure virus (680). 

The use of the electron microscope to visualize virus particles in the 
intracellular state is somewhat limited by the general opacity of cells 

Figures 29-31. Interaction between bacteriophage and bacterium as seen in the 
electron microscope. Figure 29 (left). An infected cell of Escherichia coli with 
attached particles of phagc T2. Figure 31 (right). Cell walls after lysis, sur- 
rounded by phage particles. From : Luria et al. ( 448 ) . 

Figure 30. Two infected cells during lysis, with liberation of phage particles 
and protoplasmic debris. From: Luria et al. (448). 


Size and Morphology of Virus Particles 

Figure 32. A chick red blood cell with particles of influenza virus B adsorbed 
on the cell surface. From: Heinmets (296). Courtesy Dr. F. Heinmets, Naval 
Medical Field Research Laboratory, Camp Lejeune, N. C. 

to the electron beam, due to excessive thickness. Special techniques 
have been devised for use with tissue cultures and for observation of 
very thin sections of tissues ( see chapter 11 ). The sizes of the particles 
associated with a number of viruses, as derived from electron micros- 
copy, are included in table 6 


Optical methods seldom allow direct identification of visible par- 
ticles with the active virus, since the particles observed cannot be 
tested further. The size of active virus particles, however, can often 
be estimated without inactivation even on dilute, impure virus sus- 

CH. 4 Ultrafiltration 71 

Figure 33. A droplet pattern from a mixture of tobacco mosaic virus and 
polystyrene latex particles. The concentration of the latex particles in the mixture 
is known, and the number of virus particles can be estimated from the ratio be- 
tween the counts of the two types of particles. The enlarged sector shows the 
virus particles more clearly. Note that the rod-shaped virus particles orient them- 
selves parallel to the edge of the droplet, due to the changes in surface tension 
during drying. From: Research Today, 8:28, 1952. Courtesy Dr. R. C. Williams, 
Virus Laboratory, University of California, Berkeley. 

pensions by ultrafiltration, a method based on a sieve principle. Ultra- 
filters that retain virus are valuable not only in measuring viruses but 
also in obtaining virus-free preparations and in separating different 
viruses from a mixture. 

The most useful ultrafilters are collodion membranes prepared by 
removing the solvent from thin layers of collodion solutions. The 
average pore size of the resulting collodion membrane depends on the 
concentration of collodion in the original solution, on the nature and 
the concentration of the solvent, and on the rate of solvent elimination. 
Early membranes prepared from solutions of collodion in acetic acid 
have been replaced by accurately calibrated "gradocol" membranes 
from ether-alcohol solution (204). The average pore diameter 
(A.P.D.) of a membrane is estimated indirectly from the pressure 
needed to force the first air bubbles through the membrane under 


Size and Morphology of Virus Particles 

water, or to force through it a given amount of water per unit time. 
The relation between the A.P.D. and the size of the particles that can 
go through the membrane is then obtained by filtering particles of 
known size. The correction factor, that is, the ratio between the A.P.D. 
and the diameter of the smallest particles that can pass, was originally 
given (204) as follows: A.P.D. 10-100 m/*, correction factor 3 to 2; 
A.P.D. 100-500 m/A, correction factor 2 to 1.3; A.P.D. 500-1000 m^, 


f ? 
. 5 

I 4 



*\J I I 

----- \ 

Bacteriophage h 

disease virus 


I I I 

I I I I I I 111 



200 100 25 

1000 800 600 400 

Average pore diameter, mju 

Figure 34. Filtration end-points for several viruses and other microorganisms. 
From: Elford (204). 

correction factor 1.3 to 1. According to a more recent study (474), 
however, the correction factor is about 1.25 and fairly independent of 
A.P.D. The medium in which the material to be measured is sus- 
pended plays an important role, because it affects adsorptive phe- 
nomena that take place on the walls of the pores. 

The size of virus particles is estimated from the filtration end-point, 
that is, from the A.P.D. of the finest filter that allows some virus to 
come through. For homodispersed viruses, there should be an abrupt 
transition from filters that allow all virus to pass, to filters that retain 
it all. In practice, such an abrupt change is only found for the smallest 
viruses (figure 34). The deviations for large particles depend more 
on the loss of virus particles on the walls of the pores than on in- 
homogeneity of virus materials. 

CH. 4 Hydrodynamic Methods 73 


A number of procedures for determining the size of virus particles 
rest on the analysis of the motion of the particles through the sus- 
pending fluid. Since these procedures are now widely used and many 
properties of viruses are reported in terms based on these methods, 
their principles should be clearly understood by virus students. The 
two most important types of motion to be considered are the sedi- 
mentation of virus particles in a centrifugal field and the changes in 
concentration of virus particles by diffusion. 

Sedimentation of spherical particles. The case of spherical particles 
is the simplest. A spherical particle of radius r suspended in a liquid 
will sediment in a gravitational or centrifugal field if its density d is 
greater than the density d of the medium. In a gravitational field, 
the motor force P is equal to the gravity acceleration g multiplied by 
the difference between the mass of the particle and the mass of a cor- 
responding volume of the medium: P g = %irr 3 (d do)g. In a cen- 
trifugal field, the gravitational acceleration g is replaced by the 
centrifugal acceleration c = (27ro>) 2 x, where w is the number of revo- 
lutions per unit time and x is the distance of the particle from the axis 
of rotation of the centrifuge: P c = %irr^(d d )c. 

As a particle is accelerated in a centrifuge, its velocity v increases. 
This will in turn cause an increase in friction. The frictional force is 
measured by: < vf. f is the "friction coefficient." Its value for 
spherical particles is 6^r, where ?/ is the viscosity of the medium. As 
the velocity increases, we reach a point where the frictional force 
balances the centrifugal force. The particle will then continue to 
move at a uniform velocity, which is obtained by writing P c = </>, or 


These equations express Stokes' law, which describes the motion of 
spherical particles in a fluid under ideal conditions (that is, no inter- 
action among the particles; the size of the particles large as compared 
with the size of the solvent molecules; no disturbance due to convec- 
tion currents). 

We see, then, that the motion of a spherical particle in a centrifugal 
field is determined by its size and density, by the density and viscosity 
of the medium, and by the centrifugal force. This force depends on 

74 Size and Morphology of Virus Particles 

the speed of rotation and on the size of the centrifugal rotor (or 
"head"), which determines the distance of the particle from the axis 
of rotation. From equation 6 we can derive the value of a magnitude 
S = v/c, called the sedimentation constant, which is characteristic of 
a given particle in a given medium at a given temperature. 1 Equation 
5 can then be written in the form: 

S = - = - [7] 


Equations 6 and 7 can be used to determine the radius and mass of 
a spherical particle if we have methods to measure its density and its 
velocity in a centrifuge, since the centrifugal force and the density of 
the medium and its viscosity can all be measured without difficulty. 

Diffusion of spherical particles. If a layer of a suspension of par- 
ticles and a layer of the medium alone are placed in contact along a 
boundary plane, the random motion of the particles will make some of 
them pass through the boundary. The particles will move back and 
forth, and more will cross past the plane of the initial boundary from 
the side where the concentration is higher, resulting in progressive 
equalization of the concentrations of the two sides. If the particles 
are large compared with the molecules of the medium and are few 

1 The student should remember that all units of measure for different quantities 
entering an equation must be comparable. The best method is to use the so-called 
CGS system (centimeter, gram, second). If r is expressed in centimeters, x must 
also be in centimeters, and g and c in centimeters per second. This often forces 
a change from the units most commonly employed: for example, the diameter of a 
virus particle, generally given in millimicrons, must be translated into centimeters. 
The decimal notation (see page 40) helps in handling large or small numbers: 
a radius of 12 m/t corresponds to 1.2 X 10 ~ cm, and the corresponding volume 
is %ir(1.2 X 10-0)3 cm 3 = 4.2 X 1.74 X 10~ 18 cnv* = 7.3 X 10~ cnv*. The 
value of w is often given in revolutions per minute (rpm). If w is translated into 
revolutions per second (60 rpm = 1 rps), and if x is given in centimeters, an ap- 
proximate comparison of centrifugal and gravitational accelerations is easily made, 
since the latter ( 980 cm per sec ) can be taken as about 1000, and the ratio of the 
two accelerations, that is, the ratio of centrifugal force to gravity, will be about 
1/1000 the value of the centrifugal force. 

Densities are given in grams per cubic centimeter (or in ratios of the density 
to that of water at 20 C ) . Viscosity can also be measured as the ratio to the 
viscosity of water at 20 C taken as 0.01. The sedimentation constant, that is, the 
rate of sedimentation of the particle in water at 20 C under a unity centrifugal 
force, is obtained in centimeters per second. Being generally very small, it is often 
expressed in Svedberg units (S 20 ): 1 S 20 = 10 ~ 13 cm per sec. For example, the 
sedimentation constant of tomato bushy stunt virus is 146 S 20 = 1.46 X 10 - 11 
cm per sec. 

CH. 4 Hydrodynamic Methods 75 

enough not to hamper one another, their kinetic energy will depend 
only on the temperature, and their velocity will depend on their mass 
and on the temperature. The resistance of the medium depends on 
the frictional force, that is, on the velocity of the particles and on 
their frictional coefficient (see page 73). The change in concentration 
that takes place in 1 second across 1 cm 2 of a boundary between solu- 
tions differing by 1 unit concentration is defined as the diffusion con- 
stant D. Its value is given by 

RT 9 
D = - cm 2 sec" 1 [8] 


where T is the absolute temperature, R is the so-called gas constant, 
N is the Avogadro number ( 6.02 X 10 23 ) and /, the friction coefficient, 
is given for spherical particles by &7n?r. R and T appear in equation 8 
(Einstein's equation) because of the equivalence of the motion of the 
molecules in an ideal gas and of spherical particles in an ideal solution. 
For the derivation of equation 8 the reader should consult a textbook 
of physical chemistry. 

The diffusion constant D is obtained by measuring the concentration 
of particles at different times at known distances from the plane of the 
initial boundary. 2 The radius of the particle can then be obtained 


Nonspherical particles. In the case of nonspherical particles the 
situation is complicated by the fact that the frictional force depends on 
a complex function of the linear dimensions (length, width, thickness) 
of the sedimenting particle. Measurements of sedimentation velocity 
alone or of diffusion alone do not give any direct information about the 
volume or mass of the particle. One way of getting around this 
difficulty is to measure both sedimentation and diffusion constants and 
to divide one by the other. The resulting value is independent of 

2 The actual measurements are made as follows : The suspension and the medium 
alone are placed in contact at time 0. At various intervals the concentration at 
various distances from the boundary is measured either optically or by other means. 
The concentration C x> t at a distance x and time t is related to x, t, and D by the 

-i -* 

---* dy 

76 Size and Morphology of Virus Particles 

the friction coefficient and, therefore, of the shape of the particle, and 
is proportional to the mass only: 

S m[l - (d /d)]N S RT 

m = 

D RT D N[l - 

If all units are CGS, m is obtained in grams. 

An essentially equivalent way is to measure the so-called sedimenta- 
tion equilibrium. In a relatively slow centrifugal field, the increased 
concentration of particles in the bottom of the tube will create an 
increased tendency to diffuse backward. When sedimentation and 
back diffusion balance each other, an equilibrium is reached, and from 
the concentration of particles at two given points in the centrifuge 
tube we can determine directly their mass. 

Once the mass of an asymmetrical particle is known, we can calcu- 
late what volume, radius, and friction coefficient ( f ) it would have if 
it were a sphere. The ratio of the actual value of / to the value / gives 
the "asymmetry coefficient," which indicates to what extent our par- 
ticle differs from a sphere. This criterion, however, is somewhat 
ambiguous. The same effect produced by asymmetry could also be 
produced by hydration, that is, by the particle taking up water in a 
liquid medium. A hydrated particle diffuses more slowly because its 
radius is greater* The increase in volume produced by hydration 
causes a corresponding reduction in density. Therefore, the radius 
calculated from the sedimentation velocity will be too small, if the 
value used in equation 6 for the density of the particle is derived from 
measurements on dry material. For hydrated spherical particles, if 
the density is corrected for hydration, the radius obtained from dif- 
fusion and sedimentation measurements should be the same. 

Measurement of particle density. The density of virus particles 
in the v hydrated state can itself be measured by centrifugation. If a 
particle is centrifuged in a medium of the same density as its own, no 
sedimentation takes place. The density is measured by centrifuging 
the particles in a series of media of different densities. The media 
should be solutions of substances with low osmotic activity such as 
proteins, for example, serum albumin, in order not to alter the state 
of hydration of the particles (593). The density at which the sedi- 
mentation rate would be zero is obtained by extrapolation ( figure 35 ) . 
This technique attributes to different viruses densities of 1.1-1.2 as 
against densities of 1.3 or higher obtained for dry virus preparations. 
For different viruses, the amount of hydration water varies between 
50 and 100% of the dry weight. These amounts of water are greater 

CH. 4 

Hydrodynamic Methods 


than those generally associated with purified proteins in solution 
(30-35%). This has heen taken to suggest the existence of an osmotic 
membrane at the surface of some virus particles. 

Sedimentation and diffusion apparatus. Some of the largest virus 
elementary bodies, such as those of psittacosis or vaccinia viruses, can 






; 300 

E 200 



I I r riii i \ i r 


1.00 1.02 1.04 1.06 1.08 1.10 1.12 
Solvent density 

Figure 35. The sedimentation rate of influenza virus A as a function of the 
solvent density. The density was varied by adding bovine serum albumin, to 
25%. The ordinate gives the sedimentation rate. The extrapolation to rate (no 
sedimentation) gives a density of 1.104. Data from two experiments. From: 
Sharp et al. (593). 

be sedimented in fast centrifuges of the usual laboratory type. For 
smaller viruses, high-speed centrifuges are necessary to provide cen- 
trifugal forces sufficient to sediment them. The so-called ultracentri- 
fuges, operated by electrical motors or by oil or air turbines, give 
speeds of 60,000 rpm or higher. For a particle located 6 cm from the 
rotational axis, 60,000 rpm correspond to a centrifugal force 250,000 
times gravity. Most ultracentrifuges operate in a high vacuum, thereby 
reducing friction and temperature changes. 

The construction and operation of ultracentrifuges present complex 
technical problems. For virus workers, the important problems are 

78 Size and Morphology of Virus Particles 

those connected with the measurement of the sedimentation of virus 
material. Ingenious methods have been developed to measure changes 
in concentration of virus material during centrifugation, particularly 
by workers possessing only centrifuges of rather simple construction. 
One such method consists of freezing the contents of the centrifuge 
tubes in the centrifuge head immediately after centrifugation, after 
which virus activity in various sections of the frozen columns is meas- 
ured (268). Another method (204) consists of collecting a sample 
centrifuged in a short capillary tube immersed in a larger tube. From 

Figure 36. Sedimentation diagram of purified Eastern equine encephalomyelitis 
virus in the ultracentrifuge. The photographs were taken at 2.5-minute intervals 
during centrifugation at 15,000 g. Courtesy Dr. D. G. Sharp, Department of Sur- 
gery, Duke University, Durham, N. C. 

the concentration of virus activity that remains in the capillary tube, 
the sedimentation velocity can be determined. 

With the modern ultracentrifuges, measurements are made while 
the centrifuge is in motion by means of the so-called "analytical" 
heads, in which a beam of light is sent through centrifuge tubes (or 
cells) with transparent walls. The virus preparation must possess a 
high degree of purity and must be present in high enough concen- 
tration to give measurable absorption and scattering of the light. 
As the virus material sediments, the optical density increases toward 
the bottom of the centrifuge tube and decreases at the top. If all 
particles responsible for light absorption sediment at the same speed, 
there will be a sharp boundary migrating from the top down. The 
presence of such a boundary can be determined in two ways: either by 
photographic absorption measurements through the column of liquid 
(figure 36), or by locating the boundary as a region at which the 
refractive index of the liquid column shows an abrupt change (the 
Schlieren technique). Besides giving information on the sedimenta- 
tion properties of a pure virus material, the method can also be used 
to obtain information on its degree of purity and homogeneity. A 

CH. 4 Hydrodynamic Methods 79 

suspension of identical particles should give a single boundary, whose 
sharpness is only limited by diffusion. 

With viruses ultraviolet light is often employed in the measurement 
of optical density. The high nucleic acid content of most viruses makes 
their absorption of ultraviolet light strong enough to allow analysis of 
relatively dilute solutions (0.01 mg per ml). 

Diffusion measurements are made with a variety of apparatus, all 
of which involve careful layering of a virus suspension against the 
medium alone, after which the changes in concentration at various 
distances from the boundary formed at the layering point are meas- 
ured. Optical methods similar to those employed in the ultracentrifuge 
can be used. It appears from the literature that diffusion measure- 
ments require even greater skill and are more liable to error than 
sedimentation measurements. 

It may be pointed out that the use of physical methods to determine 
changes in virus concentration requires careful and painstaking work 
to justify the identification of the material with the virus. An example 
of the results obtained by these techniques may be instructive (405). 
The tomato bushy stunt virus studied in the dry state by electron 
microscopy and x-ray diffraction has been found to have spherical 
particles of 21.8 m/x diameter. This corresponds to a mass of 13 Mh 
(density 1.3). Diffusion and sedimentation measurements indicated 
a discrepancy between the diameters calculated from diffusion constant 
and from sedimentation constant, respectively. The discrepancy could 
be explained either by a fairly high degree of asymmetry contradicted 
by the known spherical shape of the particlesor by an amount of 
hydration corresponding to 0.76 gram of water per gram of virus. 
The latter interpretation was confirmed by crystallographic measure- 
ments on crystals of virus in the wet state. The mass of the hydrated 
particle is 18 Mh units. 3 

Results of the same type were obtained with a number of spherical 
or nearly spherical particles associated with animal viruses. Sedimen- 
tation studies of sperm-shaped bacteriophages indicate the occurrence 
of a double boundary, probably due to the formation of aggregates of 
different shapes (592). Diffusion measurements on phage, when prop- 
erly made, give results in agreement with those of centrifugation (318; 
525; 539). 

Electrophoresis and light scattering. There are additional methods 
that can provide information on the size, shape, and homogeneity of 

3 As a comparison, the reader may recall that the dry mass of a small bacterial 
cell is of the order of 10 5 Mh units. 


Size and Morphology of Virus Particles 

virus particles. In electrophoresis we measure the speed of migration 
of particles in an electric field. The migration of particles depends on 
their overall electrical charge, which depends in turn on the electro- 
lytic dissociation of the acidic and basic groups. This is controlled by 
the pH of the medium. The method is useful in purifying and identify- 
ing proteins and in checking their degree of purity, and has been 
applied to several viruses (485; 539). 

Light-scattering measurements are based on the fact that a suspen- 
sion of particles scatters light from an incident beam to an extent 
depending on the number of particles and on their volume (189). 
Measurement of the amount of light scattered at different angles may 
also be used to calculate the degree of asymmetry of elongated particles. 


The illustrations in figures 15-26 and the data in table 6 give a 
representative picture of the types of elements that are the carriers of 

Figures 37-40. Polyhedra and polyhedral disease viruses from Lepidoptera. 
Figure 37 (left). Polyhedral bodies from Choristoneura fumiferana. From: Ber- 
gold, Can. J. Zool 29:17, 1951. Courtesy Dr. G. H. Bergold, Laboratory of Insect 
Pathology, Sault Sainte Marie, Ontario. Figure 38 (right). A polyhedron from 
Prodenia praefica partly dissolved in alkali, showing the polyhedral envelope and 
the virus bundles. From: Hughes, J. Bact., 39:189, 1950. Courtesy Dr. K. M. 
Hughes, University of California; Berkeley. 

virus activity in the extracellular state. These elements are generally 
called "virus particles," a term that does not commit us to any inter- 
pretation as to their nature. The particles are the objects capable of 
initiating production of more virus upon introduction into new sus- 
ceptible hosts. 

The information on virus particle size and shape does not reveal any 
basic trend or similarity among viruses as a group. The shapes vary 

CH. 4 

Size and Shape of Virus Particles 


Figure 39. A polyhedron from Porthetria dispar in dissolution. Rod-shaped 
virus particles are seen singly and in bundles. Some are within "developmental" 
membranes, which are also seen empty and collapsed. From: Bergold, in: The 
Nature of Virus Multiplication, Cambridge University Press. Courtesy Dr. G. H. 
Bergold, Laboratory of Insect Pathology, Sault Sainte Marie, Ontario. 

Figure 40. Particles from polyhedra of Bombyx mori (silkworm) in various 
stages of disintegration, showing empty "intimate membranes." From: Bergold, 
in: The Nature of Virus Multiplication, Cambridge University Press. Courtesy 
Dr. G. H. Bergold, Laboratory of Insect Pathology, Sault Sainte Marie, Ontario. 


Size and Morphology of Virus Particles 

from spheres to greatly elongated rods. Internal structures are often 
evident, possibly due to local accumulations of nucleic acid. The 
possibility of artifacts in electron microscopy should not be over- 
looked, however. Complex shapes such as tadpole-like particles, some- 
times with double tails, have been observed mainly in phages. Similar 
forms for the Newcastle disease agent and other viruses may be arti- 

Figures 41-43. Capsules and virus particles from granuloses of Lepidoptera. 
Courtesy Dr. G. H. Bergold, Laboratory of Insect Pathology, Saulte Sainte Marie, 
Ontario. Figure 41 (left). Capsules from Choristoneura fumiferana. From: 
Bergold, in: The Nature of Virus Multiplication, Cambridge University Press. 
Figure 42 (right). Virus particles from capsules of Choristoneura fumiferana. 

facts (37; 591). On the other hand, the filaments observed regularly 
in recently isolated strains of influenza (131; 494) seem to be images 
of forms really produced by infected host cells. These filaments have 
the same transverse diameter as the spherical particles of these viruses 
and give similar reactions with red blood cells. They appear to split 
into spherical particles (figure 19). Filaments described as the normal 
form of poliomyelitis virus (248) may, instead, have been artifacts. 

Among the rod-shaped particles of several plant viruses, there are 
differences in thickness, length, and flexibility. The particles of to- 
bacco mosaic virus are rods 15 m/* thick and about 300 mp, long. Their 
tendency to lengthwise and sidewise aggregation has rendered their 
study by hydrodynamic methods a most perplexing problem for the 
physicochemist (405). 

CH. 4 Size and Shape of Virus Particles 83 

A most interesting group of rod-shaped particles comprises the 
agents of the capsular and polyhedral diseases of insects ( 66; 67; 68a; 
627). The virus particles are contained in characteristic intracellular 
inclusions (figures 37-43). In capsular diseases (granuloses) the 
capsules or granules are mainly intracytoplasmic. In polyhedral dis- 
eases, the polyhedra are intranuclear. The virus particles are rigid 
rods, mostly around 40 X 300 m/x, somewhat larger for some of the 
viruses (627). The rods are present singly in each capsule. In the 

Figure 43. Empty capsules from Cacoecia murinana. 

polyhedra they are in bundles (figures 37-40). Each polyhedron con- 
tains several hundred virus particles, contained in a protein matrix 
with a distinct membrane. 

Other particles of varying morphology are also found in the poly- 
hedra, and have been used in reconstructing a developmental cycle 
of the rods from small spherical elements (68; 68a; figures 39, 40). 

The individual insect virus particles possess an intimate membrane 
and seem to be formed in bundles within a common developmental 
membrane. These details are best observed in polyhedra that have 
been partially disintegrated with alkali. In some diseases the poly- 
hedra contain mostly spherical particles, only exceptionally rod par- 
ticles (611). 

There are obvious limitations to the usefulness of information on the 
morphology of extracellular forms in the study of intracellular para- 
sites. Increasing evidence is accumulating to show that viruses during 
their reproduction in the intracellular state differ in morphology and 
organization from the free virus particles, which may be thought of 

84 Size and Morphology of Virus Particles 

as resting stages of microorganisms that go through complex life cycles 
(41; 439). Even in the present rather primitive state of our knowledge 
on the organization of virus particles, several facts suggest that many 
of them differ from the reproductive form of the virus by possessing 
some kind of protecting covering or membrane. This is clearly so 
with some bacteriophages. Upon infection of a bacterium, the pro- 
tein envelope of the phage is left on the surface of the host cell (313), 
and new protein envelopes are formed late in the process of maturation 
of the new particles (175; see chapter 8). 

Among other viruses, vaccinia particles seem to have a membrane. 
The capsules of some insect viruses may have a protective function 
(see 355a). Any such coverings would have to function also as a 
device for bringing or releasing the reproductive element of the virus 
into its new host cell. Further discussion of these ideas will be found 
in chapters 8-13. 



Purification and Chemical Composition 
oi Virus Material 


The material that carries virus activity in crude preparations is 
mixed with large amounts of foreign matter. Before any study on its 
chemical composition can be made, the specific virus material must 
be separated from the extraneous materials. We can define the goal 
of purification as the preparation of material carrying virus activity 
in a form as free as possible from iioninfectious matter. Immediately 
we meet with theoretical difficulties, many of which have been pointed 
out by Pirie ( 520 ) . Any method of purification will depend upon what 
we consider as pure. Criteria of purity are by no means definite. They 
vary with the interest of the investigator. An analytical chemist would 
aim at obtaining a virus preparation consisting of particles identical in 
size and chemical composition, each carrying equal and full activity 
and representing a "molecule" of virus, according to the usual defini- 
tion of molecule as the minimum amount of matter endowed with all 
the chemical properties of a substance. A physicochemist, by the use 
of his techniques, would tend to consider as pure a population of 
particles of uniform size that also satisfy additional criteria, for 
example, the possession of uniform numbers of acid and of basic groups 
as measured by electrophoresis. A biologist would probably consider 
as ideal a purification method that could separate all particles with 
virus activity ( ability to reproduce ) from all inactive material. 

None of these criteria is either clear-cut or operationally satisfactory. 
The uniformity of chemical composition may fail both as a theoretical 
criterion since a variety of chemical forms and configurations may be 
compatible with a given virus activity and as a practical criterion 
since the gross chemical composition of most proteins is very similar 
and their finer organization is still beyond our means of study. For 


86 Chemical Composition of Virus Material 

example, the chemical differences between active virus particles and 
particles inactivated by mild treatments are generally well beyond the 
analytical techniques of the most skillful chemist. The same can be 
said of the purely physicochemical criteria, which force upon virus 
material requirements for uniformity in size, shape, or electric charge, 
while neglecting more relevant information. For example, a mixture 
of active and radiation-inactivated tomato bushy stunt virus crystallizes 
in exactly the same way as a "pure" preparation of active virus. Is the 
mixture still "pure"? 

The biological criterion of activity may also be deceiving, both 
practically, by requiring an often impossible separation of active from 
inactive particles, and theoretically, by postulating that association of 
activity with a given material identifies all this material with the virus 
and that all inactive material is nonvirus. We should remember that 
the definition itself of the minimum amount of material required to 
initiate virus production presents several ambiguities. On the one 
hand, the "true virus" might be adsorbed on inert carriers, still re- 
movable without impairing virus activity. On the other hand, the 
elimination of all material not essential for virus infectivity, as tested 
by reinoculation, may lead to the neglect of some important aspect of 
the virus. In fact, there is no reason to assume that only what is 
strictly necessary for infectivity is virus. By similar reasoning, we 
should say, on the basis of experiments on artificial insemination, that 
the only material that can be defined as a male animal is its sperm. 
In other words, the criterion of infectivity ability to enter a sensitive 
host and to reproduce in it may be too restrictive, since it may ignore 
specific portions of the virus which are connected with virus activity 
only when the virus is in its intracellular state (50; 522). 

Some of these problems may appear to be peculiar to viruses, but 
we must remember that the question of purity, interpreted in chemical 
terms, is raised for viruses and not for full-fledged organisms, because 
the small size and presumably simple organization of virus particles 
suggest for them a simpler, more easily interpretable chemical structure. 

If we decide that all criteria of purity must be operational, that is, 
defined in terms of meaningful, performable operations, we can state 
the problem of the physical and chemical structure of virus particles 
in the following form: Is a certain material, present in a virus sus- 
pension, specific? If so, what are its chemical and physical properties? 
Is its presence necessary for virus transmission and reproduction? Is 
this material related to specific, noninfectious materials present in the 
same suspension? Is the infectious material present in one form only 

CH. 5 The Meaning of Purity 87 

(homogeneously dispersed, chemically uniform)? What components 
of the suspension are necessary or responsible for the various phases of 
the virus interaction with the host? Is there any detectable relation- 
ship between the specific material and some nonspecific component of 
the host? 

In these general terms, the problems of virus purity cover practi- 
cally the entire field of virology. This formulation, however, gives us 
some guidance in the more modest task of analyzing the properties 
of those particles with which it appears reasonable to identify, at least 
in first approximation, the specific virus material in infectious prepara- 
tions. It has been pointed out that the very process of extraction can 
modify virus material. For example, viruses affecting strawberry plants 
can hardly be studied in vitro in strawberry extracts because of the 
presence in the plant sap of enough tannic acid to precipitate not only 
the virus but also most protein materials (47). Another case in point 
is that of pneumonia virus of mice, which, if extracted by grinding the 
infected lung tissue, becomes associated with a nonspecific host com- 
ponent (160). 

When virus activity has been attributed on reasonable grounds to 
certain material particles, these can be separated and their chemical 
composition can be analyzed. The process of purification should be 
accompanied by an increase of virus activity per unit mass of the 
purified material. This is not always so, however, because of oc- 
casional losses of activity in the course of the purification process, 
either by inactivation, or by irreversible aggregation, or by loss of a 
portion of activity that was associated with other materials. These 
losses sometimes make it difficult to use the results of purification pro- 
cedures as evidence for the identification of virus activity with one or 
another type of particles. Since all viruses have been found to consist 
in large part of protein material, a convenient way of following the 
decrease in foreign material in the course of purification is to measure 
at different stages of the process the number of infectious units per 
unit weight of protein nitrogen. An increase in virus titer per milli- 
gram of protein nitrogen indicates a removal of noninfectious proteins. 

Activity measurements on purified virus preparations, although giv- 
ing some indication as to the extent to which impurities have been 
eliminated, can seldom be considered as a test of purity, because of 
the low efficiency of most titration methods, as discussed in chapter 3. 
The best preparations of tobacco mosaic virus, for example, are infec- 
tious in concentrations of 10~ 13 to 10 " u gram of protein per ml. This 
still represents over 1000 particles 40 Mh in weight. Still, suspensions 

88 Chemical Composition of Virus Material 

with such a low virus content do not give any of the usual tests for 
proteins, which at best are sensitive to concentrations of 10 r ' gram 
per ml. This explains, without justifying them, early claims for "pro- 
tein-free virus preparations." 

Small amounts of impurities can foe detected in purified preparations 
of plant viruses by the very sensitive serological tests. Chester (130) 
found that the uterine horn of a guinea pig sensitized with normal plant 
sap would respond with an anaphylactic contraction to the addition of 
purified tobacco mosaic virus preparations to the surrounding fluid 
(Shultz-Dale technique). In some cases, the impurities could be re- 
moved by treatment with proteolytic enzymes, which do not attack 
the virus. Bacterial impurities in purified phage preparations can be 
removed by precipitation with antibacterial serum (146). 


The methods employed for the separation of viruses are borrowed 
from protein chemistry. We can employ chemical treatments such as 
"salting-out" procedures, which consist of the addition of high concen- 
trations of very soluble salts, for example, ammonium sulfate (neutral- 
ized). Proteins are forced out of solution at salt concentrations that 
fairly well characterize different groups of them. For example, serum 
albumin is defined as a serum fraction that precipitates at 100% satura- 
tion with ammonium sulfate, whereas serum globulins precipitate at 
50$ saturation. It is also possible to precipitate viruses by adding 
alcohol or by adjusting the pH of the medium to the point of minimum 
solubility (isoelectric point). Both types of treatment can be used in 
the purification of tobacco mosaic virus, which precipitates at % to 
a /2 saturation with ammonium sulfate and is stable at its isoelectric 
point (pH 3.4). Some viruses, for example, those of foot-and-mouth 
disease and of poliomyelitis, have been separated by first adsorbing 
them with colloidal precipitates of Al(OH) a or CaSO 4 and then eluting 
them. Large virus particles can be precipitated by treatment with 
protamine sulfate, which leaves in suspension particles 50 m/x or less 
in diameter (664). 

Physical methods of virus separation are generally based on size 
differences between the virus particles and other materials in suspen- 
sion. Centrifugation is used in a series of cycles alternating low-speed 
centrifugation, which removes tissue debris or bacterial contaminants, 
and high-speed centrifugation, which collects the virus in a pellet. 
Some virus particles, such as the elementary bodies of vaccinia, can be 
sedimented in fast centrifuges of the ordinary type; alternate cycles at 

CH. 5 Methods of Purification 89 

3000 rpm in a horizontal centrifuge and at 4500 rpm in an angle centri- 
fuge yield fairly pure suspensions of elementary bodies (155). For 
smaller viruses, faster centrifuges can be used, equipped with so-called 
"quantity heads" that carry several large centrifuge tubes. The 
Sharpies centrifuge, in which the material is sedimented from a thin 
layer of fluid covering the inner surface of a rotating hollow cylinder, 
is also suitable for virus purification. A procedure that promises to be 
very useful is "density gradient centrifugation" (lOOa). The virus 
particles are concentrated in specific layers, as they sediment in a 
medium with a density gradient from top to bottom. 

All these methods allow only separation of viruses which differ in 
size from nonviral components of the crude suspensions. These 
methods may cause the loss of some virus, if the virus consists of a 
variety of particles of very different sizes. Control of the activity in 
each discarded fraction during a purification procedure should avoid 
the overlooking of any important active fraction, but may not reveal 
small amounts. A certain loss of activity always occurs, particularly in 
the centrifuge, because of incomplete sedimentation. For example, in 
vaccinia virus, centrifugation always leaves in the supernatant some 
activity, which has been shown to be accounted for by elementary 

Some procedures employed in virus 'purification are based on 
peculiarities of the individual viruses themselves. Tjpeatment with 
proteolytic enzymes, as employed in purification of several viruses, is 
based on the fact that the virus itself is not attacked as long as it is 
active. Influenza virus particles can be purified by allowing them to 
become specifically adsorbed onto red blood cells at low temperature. 
The cells are collected, and upon incubation at 37 C there is elution 
of the virus, which can then be easily separated from the cells by 
centrifugation (233). Influenza viruses can also be purified by ad- 
sorption and elution on cation exchange resins (496a). 

The purified material obtained by various procedures can be sub- 
mitted to a number of tests to study its properties and its relation to 
virus activity. An important phase of this study is the comparison with 
materials obtained by similar procedures from uninfected hosts. A step 
forward in the identification is the proof that an identical material can 
be isolated from different hosts infected with the same virus, whereas 
no similar material can be Obtained from uninfected hosts. At the 
same time, a comparison of the properties of the purified virus with 
those of normal host components may provide some lead to the mode 
of virus production. 


Chemical Composition of Virus Material 


Physicochemical analysis. Some plant viruses are particularly suit- 
able for purification studies, because the specific materials can be 
obtained in large amounts from infected plants (up to IQ% of the dry 

Figure 44. Dodecahedral crystals of purified tomato bushy stunt virus. From: 
Bawden and Pirie, Br. J. Exp. Path. 19:251, 1938. Courtesy Dr. F. C. Bawden, 
Rothamsted Experimental Station, Harpenden, England. 

weight of the plants for tobacco mosaic virus). These materials can 
often be purified without loss of activity, to yield protein preparations 
as homogeneous as those of the best-known proteins. A number of 
these "virus proteins" are obtained in beautiful crystalline forms. 
Tomato bushy stunt virus gives uniform rhombic dodecahedral crystals 
(figure 44); tobacco necrosis virus yields a variety of regular crystal- 

CH. 5 

Purified Virus Preparations 


line forms. Southern bean mosaic virus, turnip yellow mosaic virus 
(figure 45) and squash mosaic virus also give fairly perfect crystals. 
Some of these crystals have been shown ( see 688 ) by electron-micro- 
scopic study of their surfaces to be made up of very orderly arrays 
of spherical or nearly spherical particles of great uniformity (figure 

Figure 45. 
Smith (473). 

Crystals of turnip yellow mosaic virus. From: Markham and 
Courtesy Dr. K. M. Smith, Molteno Institute, University of Cam- 

46). The size of the particles in the crystals can be estimated by 
crystallographic methods such as x-ray diffraction analysis, by which 
the distance between successive planes of symmetry within a crystal 
structure is calculated (475). A comparison of dry and wet crystals 
of tomato bushy stunt virus (70) has indicated that upon hydration the 
distance between the centers of individual particles increases by an 
amount corresponding to about 70% hydration by weight, in agree- 
ment with the conclusions derived from sedimentation and diffusion 

92 Chemical Composition of Virus Material 

The formation of regular crystals has been claimed as evidence for 
both the purity and the simplicity of constitution of these crystallizable 
virus proteins. As far as purity is concerned, crystal formation is a 
good but not an absolute criterion. The possibility exists that foreign 
materials may gain seme degree of hospitality within a crystal, In- 
active bushy stunt virus crystallizes just as well as active virus, 
although its chemical structure must be somewhat altered. Moreover, 

Figure 46. A crystal of tobacco necrosis virus. Courtesy Dr. K. M. Smith, 
Molteno Institute, University of Cambridge. Photographs from the same study 
were published by Markham ct al. (475). 

the_bushy stunt virus in combination with substances such as clupein 
(a protamine) or heparin also crystallizes in regular forms. 

A second group of plant viruses, including potato yellow dwarf and 
tomato spotted wilt, consists of particles about 100 m/u in diameter, 
whose electron-microscopic appearance is that of flattened spheres or 
ellipsoids (lOOb). These viruses have not been crystallized. 

A third/group of plant viruses yields purified preparations (622) 
consisting of very elongated particles. This group includes, among 
others, tobacco mosaic virusj some cucumber viruses, and potato 
viruses X and Y ( see figures 20, 21 ) . The particles of the potato viruses 
are apparently less rigid and more flexuous than the tobacco mosaic 
virus rods (688). The physical properties of purified preparations of 
these viruses reflect the asymmetry of the individual particles. 

CH. 5 Purified Virus Preparations 93 

"fobacco mosaic virus is probably the virus whose purified prepara- 
tions have been most thoroughly investigated and have raised not 
only complex physical problems but also heated polemics among virolo- 
gists. ^Preparations from plants infected with tobacco mosaic virus 
were shown by Stanley in 1935 (622) to yield a highly infectious pre- 
cipitate consisting of microscopic needles, which Stanley considered 
to be crystals of virus protein, wm auusequent work has confirmed the 
identification of the needles with the specific virus nucleoprotein and 
has shown that the needles are highly organized bundles of the rod- 
shaped virus particles regularly arranged side by side (69). 

(Lauffer (403) lists the following evidence for the identification of 
the tobacco mosaic virus protein with the virus itself: constant asso- 
ciation with virus activity; independence of the host on which the 
virus is grown; similarity among the proteins corresponding to related 
virus strains; similarity between the ultraviolet absorption spectrum of 
the virus and the action spectrum for virus inactivation; equal pH 
stability range for virus activity and protein homogeneity; similar sedi- 
mentation rates for the infectious material (measured in the separa- 
tion cell of the ultracentrifuge) and the protein. 

Electron micrographs of tobacco mosaic virus material always con- 
tain a variety of particles of very uniform thickness (about 15 m/x) 
but of somewhat variable length, from 40 to 500 m/x or more, with a 
mode around 290-300 m/x ( 688; figure 20 ) . This observation has sug- 
gested that the protein may be present in the form of small units in 
different states of linear aggregation!) This raised the question as to 
what the true "virus particle" and its real significance may be. Some 
workers (see 43) have pointed out that infected plants yield some 
virus protein in the form of less asymmetric elements than the long 
particles visible in electron micrographs. They suggested that the 
virus in the native state consists of elements shorter than 300 m/x, which 
aggregate linearly upon extraction. Stanley and several other workers, 
however, have shown that infectivity is never present in preparations 
which do not contain particles of approximately 280-300 m/x, and con- 
sider these elements as the infectious particles (507). 

The best purified preparations sediment in the ultracentrifuge (in 
albumin solutions) with a sharp single boundary corresponding to a 
sedimentation constant of 230S , suggesting homogeneity of size and 
a particle size about 15 X 300 m/x (405). The presence of a variety 
of particle lengths in electron micrographs has been attributed to 
breakage or aggregation in the drying process. In spray-drop patterns, 
in which a fine mist of virus suspension is sprayed onto electron-micro- 

94 Chemical Composition of Virus Material 

scope specimen holders and the patterns formed by individual droplets 
are photographed, practically all the virus fragments that differ from 
the typical 300-m/x unit in any given individual pattern add up to this 
basic unit length or to its small multiples. This suggests that they are 
produced within the droplets by breakage of units of the 300-m//, 
length or of dimers or trimers of them (678). 

There is certainly no doubt that the virus protein is the virus ma- 
terial, in the sense that the virus activity is embodied in particles 
consisting of the protein. There is also little doubt of the constant 
association of virus activity with the "Stanley unit" ( 15 X 300 m/x par- 
ticle). The objections raised to these conclusions have served, how- 
ever, to focus attention on the fact that the virus protein and its 
component units may provide information not only on the nature of 
the infectious virus particles but also on the organization and produc- 
tion of virus inside the host cells. 

(Preparations of tobacco mosaic virus and other rod-shaped plant 
viruses exhibit a number of interesting properties, reflecting the tend- 
ency of the rod-shaped particles to orient themselves in parallel and 
to organize into bundles J Flow birefringence (641) results from the 
grating action exerted on polarized light by highly asymmetrical par- 
ticles that become oriented in a moving column of fluid. Concentrated! 
suspensions of Jtfirus protejp tend to separate into two layers (layering 
phenomenon; 43), the bottom layer being a more concentrated and 
more rigidly- oriented solution, representing what is called a liquid 
crystal In the liquid phase, the rod-shaped particles orient themselves 
sidewise into "tactoids," which then settle into the liquid crystalline 
phase. Microscopic tactoids can actually be seen at the border be- 
tween the two layers by examination in polarized light (figure 47). 
Bernal and Fankuchen (69) have shown by x-ray diffraction analysis 
that the crystalline needles obtained by precipitation are similar in or- 
ganization to the tactoids of the liquid crystal phase. In the needles, 
which should more correctly be called "paracrystals," the individual 
rods are oriented sidewise with great regularity, but are not spaced 
regularly lengthwise. Indeed, the needles may not be produced as 
such by accumulation of particles, but may represent isolated portions 
of a more extensive gel of liquid crystalline properties. 

The x-ray diffraction studies have provided further interesting infor- 
mation (69). In dry paracrystals, the centers of adjacent particles lie 
15.2 m/A apart, a value close to the thickness of the virus particles as 
shown by electron microscopy. Upon wetting, the dry crystals take up 

CH. 5 Purified Virus .Preparations 95 

water, and the sidewise spacings increase to as much as 23 m/x before 
the paracrystals show signs of disorientation. The persistence of the 
spatial relation between wet particles separated by several layers of 
water molecules has suggested that the particles are kept together by 
long-range forces. These need not be specific forces acting at a dis- 

Figure 47. The interface between top and bottom layers of a concentrated 
solution of tobacco mosaic virus, photographed in polarized light. Birefringent, 
spindle-shaped droplets ( tactoids ) form in the upper layer and settle in the bottom 
layer. From: Bawden and Pirie, Proc. Roy. Soc. (London), B, 123:274, 1937. 
Courtesy Dr. F. C. Bawden, Rothamsted Experimental Station, Harpenden, Eng- 

tance (no such physical force having been proved to exist). Bundles 
of oriented particles of great length may manage to retain between one 
another more layers of oriented water molecules than ordinary pro- 
teins, so that the oriented water layers provide stabilizing forces for 
the bundles. It is also possible that a few points of contact among the 
long particles may be sufficient to preserve their orientation in the wet 

Both x-ray diffraction studies (69) and high-resolution electron 
microscopy (676a) indicate that the individual virus rods have a 
hexagonal cross-section. Concerning the internal structure of the 
particles, x-ray diffraction analysis has suggested the occurrence of 
regular subunits. A particle would consist of a repeated series of 
subunits piled on top of one another like coins in a roll. 

96. Chemical Composition of Virus Material 

/The ^particles of tobacco mosaic virus can be disintegrated in vitro 
by mild alkaline treatment into progressively smaller fragments; it has 
been claimed that each class of fragments possesses a high degree of 
homogeneity (587). The fragments have been supposed to originate 
by regular cleavage of successive subunits or groups of subunits from a 
native particle r !5 X 300 m/n. j All fragments seem to carry the full 
serological specificity of the virus, although some of the smallest ones 
lack the nucleic acid component. None of the fragments possesses 
infectivity. The chemical integrity of the fragments was supposedly 
confirmed by electrophoretic analyses, which indicated that the ratio 
of free acid to basic groups was the same in the fragments as in the 
large particles. The remarkable claim was made that upon mild acid 
treatment the fragments could come together again and reconstitute 
a particle with the physical properties of the original one, although 
inactive. This work requires confirmation. 

The suggestion that a virus such as tobacco mosaic virus consists 
of repeated subunits implies that the specific element, which is repli- 
cated in the course of intracellular reproduction, is the individual sub- 
unit rather than the particle as a whole. Such a conclusion requires 
subsidiary hypotheses to explain why the infectivity, that is, the ability 
to initiate reproduction upon inoculation, is limited to the specific level 
of polymerization corresponding to the Stanley unit. 

In virus-infected cells are found hexagonal crystals which have been 
thought to consist of protein elements smaller than the infectious virus 
particles. Recently, however, such crystals have been isolated by 
micromanipulation from frozen-dried cells and have been shown to 
contain almost exclusively the typical rod particles (625a; see frontis- 
piece). These intracellular crystals may consist of layers composed 
of full-length virus particles standing side by side ( 676 ) . 

Chemical analysis. The chemical analysis of many purified viruses 
reveals differences among individual viruses and groups of viruses, dif- 
ferences which, although as yet insufficient to allow a chemical inter- 
pretation of virus specificity and reproduction, have helped placing the 
viruses in their proper company among biological objects. In all 
viruses, a high content of protein is found to be accompanied by the 
presence of nucleic acid. All viruses may be said to contain nucleo- 
proteins. Purified plant viruses have been found to yield only protein, 
nucleic acid, and small amounts of ash, indicating a very simple com- 
position. Moreover, the protein component for each plant virus ap- 
pears chemically to be all of one type (405). 

CJH. 5 Purified Virus Preparations 97 

\Nucleic acids (280) consist of combinations of nucleotides, each 
nucleotide containing a purine or pyrimidine base linked to a pentose 
molecule, which in turn is linked to phosphoric acid by its C atom 
in position 5. Successive nucleotides are held together by phospho- 
ester bonds between the phosphoric acid residue of one nucleotide and 
the pentose of the next one, at the carbon atom in position 3.) The pen- 
tose is found to be ribose in the nucleotides of nucleic acid from yeast 
and desoxyribose in nucleic acid extracted from the thymus gland. 
Every nucleic acid fraction extracted from any one source has been 
found to contain either a pentose or a desoxypentose. ) The correspond- 
ing sugars are assumed to be ribose and desoxyrioose, respectively 
( "ribonucleic acids" or RNA and "desoxyribonucleic acids" or DNA). 
Chemical and microchemical methods to distinguish between the two 
types of nucleic acid are available, based on the stronger aldehyde 
function of the desoxypentose than of the pentose? The Feulgen test 
for DNA, widely used in cytochemistry, is based on this difference^) 
The purine and pyrimidine bases isolated from a given nucleic acid 
preparation consist generally of two purines and two pyrimidines, so 
that each nucleic acid contains four types of nucleotides./ Some nu- 
cleic acid preparations contain five rather than four bases (471). /The 
possibility for a variety of specific nucleic acid structures results from 
differences in the nature and sedation of the bases )( 128; 687a).( The 
purines are adenine and guanine; the pyrimidines are cytosine and 
uracil in RNA, cytosine and thymine in DNA. Cytosine is sometimes 
accompanied and partly replaced by methyl-cytosine ( although not in 
viruses); in one group of bacteriophages (see chapter 8) cytosine is 
replaced by hydroxymethyl-cytosine. (The molar ratios (adenine: 
thymine) and ( guanine : cytosine ) are always close to unity in various 
DNA preparations; the ratio (adenine : guanine ) varies rather widely 
in DNA from different sources. , 

(The purine and pyrimidine rings of nucleic acids are responsible for 
the characteristic absorption spectrum of these compounds in the ultra- 
violet range, with a minimum at 2500 A, a maximum around 2600 A, 
and a rapidly decreasing absorption between 2600 and 3000 Aj( figure 
48 ) . The presence and intensity of the 2600- A maximum can be used 
for approximate quantitative determinations of nucleic acid. The high 
phosphorus content of nucleic acids (approximately 10%) accounts in 
part for the relative opacity of nucleoproteins in electron micrographs. 
An important reaction of nucleic acids is their depolymerization by 
specific enzymes or nucleuses. Two such enzymes have been isolated 
from pancreas, a ribonuclease and a desoxyribonuclease. The fact 
that all known nucleic acids (after separation from the conjugated 

98 Chemical Composition of Virus Material 

proteins) are split by either one or the other enzyme confirms the 
hypothesis that all nucleic acids belong to one of two groups, differ- 
entiated by the pentose they contain. ] 

The protein components of the viruses appear to share the structure 
of proteins in general (294); they consist of amino acids joined into 







260 280 300 

Wavelength, millimicrons 

Figure -48. Ultraviolet absorption spectra, (a) Bacteriophage T2. (b) To- 
bacco mosaic virus, (c) Yeast nucleic acid, (d) Serum globulin. The optical 
density in the wavelength region beyond 320 m/* is due to light scattering. Note 
the resemblance between the spectrum of the nucleic acid and that of the phage, 
which contains about 40% nucleic acid. In the spectrum of tobacco mosaic virus, 
which contains 94% protein, the absorption due to proteins is more evident. 

polypeptide chains by peptide bonds. From the peptide-chain skele- 
ton depart side chains representing the individual amino acid residues, 
whose nature and proportions give a protein its specific composition. 
1 Proteolytic enzymes split proteins by hydrolysis of peptide bonds, the 
specific site of attack of an enzyme (for example, pepsin or trypsin) 
being regulated by the nature of the adjacent amino acid residues. The 
characteristics of the surface groups determine a number of the prop- 
erties of a protein. Affinity for water is regulated by the amount of 
polar hydrophilic groups. The acidic or basic character of a protein 
; depends on the ratio of acidic residues, such as those of the dicar- 
I boxylic amino acids, to the basic residues of arginine, histidine, and 
1 lysine. The aromatic amino acids (tryptophan, tyrosine, phenyl- 

CH. 5 Purified Virus Preparations yy 

alanine) give proteins a characteristic ultraviolet absorption witn a 
maximum around 2800 A (figure 48). The organization of the pep- 
tide chain is well known only for certain fibrous proteins, along which 
the successive amino-acid residues have been found by x-ray analysis 
to be ranged at distances of 3.3 to 3.5 A. Most proteins, however, are 
in globular form, with the polypeptide chains folded in complex ways. 

In summary, our knowledge of protein (and nucleic acid) structure 
is too inadequate to permit any attempt to interpret the biological 
activities of a protein or nucleoprotein in terms of chemical structure. 
The specificity of a protein depends at least as much on the fine details 
of the folding of its polypeptide chains and on the resulting peculiarities 
of its reactive surface as on its amino acid composition. The role of 
either level of structural organization in determining specific biological 
properties is practically unknown. Very slight changes in configuration 
of a folded peptide chain might be compatible with uniformity of 
physical and chemical characteristics but not with biological activity. 
Such slight differences in configuration may be involved, for example, 
in the transformation of certain inactive enzyme precursors into the 
corresponding active forms (499). 

A number of proteins are associated with fairly simple nonprotein 
groups (prosthetic groups), which sometimes consist of metal ions and 
sometimes of organic compounds. In proteins with enzymatic activity, 
the prosthetic groups may be necessary for this activity. There is no 
evidence whether the relation of nucleic acids to proteins resembles 
that of prosthetic groups. The linkage of nucleic acid to protein is 
still a matter of controversy. Salt linkages between the acidic and 
basic groups of the nucleic acid and of the proteins, respectively, are 
probably involved. It is even doubtful that we are justified in speak- 
ing of "nucleoproteins" as chemical entities. The association of nucleic 
acids with proteins is a matter of joint extractability from natural 

In spite of our ignorance of the chemical function of nucleoproteins, 
some ideas on their possible biological role have been derived from a 
knowledge of their distribution and relative amounts in living matter 
(150). Thus, nucleoproteins are found wherever we have reason to 
suspect the presence of the determinants of specificity for newly formed 
biological material: in the cell nucleus, in the self -reproducing cyto- 
plasmic units such as the chloroplasts of plant cells, and in the viruses. 
Moreover, nucleoproteins are probably involved in protein synthesis. 
They increase in amount in cells actively involved in synthetic 
processes. Desoxyribonucleic acid is generally found only in nuclei, 

100 Chemical Composition of Virus Material 

whereas ribonucleic acid is more common in the cytoplasmic con- 
stituents of most cells. Speculations on the role of nucleic acids in- 
clude the possibility that the orderly array of phosphate residues serves 
to channel the energy required by synthetic processes in the geometri- 
cally organized fashion that may be required for synthesis of large 
specific molecules. This suggestion is based on analogy with the role 
of various nucleotides as coenzymes in hydrogen transfer. It has also 
been suggested (293) that the nucleic acids may supply the rigid 
framework required for the maintenance of the configuration of a 
protein in the unfolded, two-dimensional state, in which a protein 
molecule, acting as a model for production of similar molecules, might 
have to arrange itself in order to make possible an identical, point-by- 
point replication. Finally, it seems likely that the nucleic acids may 
themselves be, in part or exclusively, the carriers of specific biological 
configurations (32). 

Watson and Crick (666a) have proposed for DNA a structure con- 
sisting of two helical polynucleotide chains coiled around the same 
axis and held together by bonds between the purine and pyrimidine 
bases. This structure accounts for the molar ratios of various bases 
in DNA and suggests a mechanism by which specific sequences of 
nucleotides can be duplicated exactly. Thus, the proposed DNA 
structure would -make possible the reduplication of specific genetic 
properties embodied in DNA organization. 


All the nucleic acids isolated from plant viruses resemble yeast 
nucleic acid in containing only ribose. Some differences exist between 
the purines and pyrimidines of tobacco mosaic and of yeast nucleic 
acids ( 472 ) . The amount of nucleic acid varies greatly among different 
virus proteins, as shown in table 7. The amounts of purines and 
pyrimidines isolated from different plant viruses differ (table 8), the 
more closely related viruses being more similar in composition (471). 
It is likely that each virus will be found to contain a variety of nucleic 
acid elements of different structure. 

The size of the nucleic acid "molecules" isolated from plant viruses 
depends on the mode of isolation. Tobacco mosaic nucleic acid, if 
extracted by alkali, is in the form of small units (about 11,000 mol. 
wt.); if extracted by rapid heat treatment it is in the form of long 
filaments (300,000 mol. wt.). These have been supposed to represent 
the native form of. nucleic acid, possibly holding together small blocks 

CH. 5 Plant Viruses 101 

Table 7. The nucleic acid content of plant virus proteins 

From Markham (471) 


Tobacco mosaic and strains 
Tomato bushy stunt 
Tobacco necrosis 
Southern bean mosaic 
Potato X 
Turnip yellow mosaic 

Per Cent 





Table 8. Purine and pyrimidine composition of the nucleic acids 
from plant viruses related to tobacco mosaic 

From Markham and Smith 

(The molar proportions of the four bases (mean = 1) together with their standard errors are given. 
The figures in italics differ significantly from the corresponding values for tomato mosaic virus nucleic 








of Deter- 


Aucuba mosaic 






Type tobacco mosaic 

1.24 db 0.026 

1.17 0.023 

0.62 0.021 

0.96 0.019 








Cucumber 4 

1.04 0.009 

1.03 0.009 

0.74 0.006 

1.19 0.012 


Tomato mosaic 

1.18 0.007 

1.04 0.014 

0.73 0.004 

1.05 0.037 


Mean of all strains 






Yeast nucleic acid 






of protein in the elongated particles (148). In view of the ready 
tendency of nucleic acids to aggregate in vitro into long fibers, the 
heat-extracted filaments might be artifacts. Treatment of tobacco 
mosaic with alkali has suggested that the nucleic acid may belong in 
the smallest fragments derived from the surfaces of the particles ( 587 ) . 
For turnip yellow mosaic virus, it has been suggested (473) that the 
nucleic acid is contained inside a shell of protein. 

The purine and pyrimidine components and the amino acid com- 
position have been studied for several plant viruses (table 9). Inter- 


Chemical Composition of Virus Material 

Table 9. Amino acid content of highly purified preparations of some 
strains of tobacco mosaic virus * 

From Knight (380) 

TMV tobacco mosaic virus. M = Holmes' masked strain. J14D1 = a variant of TMV. GA = 

green aucuba mosaic virus. YA yellow aucuba. HR - Holmes' ribgrass strain. CV3 and CV4 

cucumber mosaic viruses 3 and 4. 

Amino Acid 
























11 .1 






Aspartic acid 

















Glutamic acid 





























































































































* The values given in the table represent percentages of the indicated amino acids. In order to 
facilitate comparison, the values which are considered to differ significantly from those of TMV are 
in italics. 

f Mean deviation of the values of single determinations from the averages given. 

esting differences are correlated with the presumed degree tif genetic 
relationship (380). Distant relatives, such as tobacco mosaic and 
Holmes' ribgrass virus, differ in a number of respects, whereas more 
closely related viruses differ in their content of only one or two amino 
acids or in the proportions of some aromatic bases. The differences 
represent changes of several hundred or thousand residues per virus 
particle. This again suggests a repeat structure for these viruses. 
Differences in amino acid composition are also reflected in differences 
in electrophoretic behavior among members of the tobacco mosaic 
group (243). Interestingly enough, the proteins extracted from many 
viruses are acidic, not basic like those found associated with nucleic 
acids in nucleoproteins extracted from most cells (protamines and his- 

CH. 5 Bacteriophages 103 

tones ) . Protein and nucleic acid also appear to be more firmly bound 
in viruses than in cellular nucleoproteins (see 381). Cucumber mosaic 
viruses 3 and 4 seem to contain no sulfur amino acids; absence of sulfur 
is clearly not adequate evidence for absence of protein. 

A number of workers have produced artificial chemical modifications 
in tobacco mosaic (624). Rather extensive chemical changes can be 
made without impairing virus activity. For example, 70% of the free 
amino groups can be blocked with acetyl radicals without loss of 
activity, and all SH groups can be covered with iodine. Further 
iodine treatment, causing the transformation of tyrosine residues into 
diiodotyrosine, produces inactivation. In all cases, new virus produced 
by inoculation of the modified but active forms into plants is of the 
normal type. In view of the improbability that the host plants may 
possess enzymes that remove the covering groups prior to reproduc- 
tion, Stanley suggested that only certain parts of the structural con- 
figuration of the virus particle are necessary to initiate the reproductive 
process. It is possible that only the viral nucleic acid is involved in the 
reproductive function. Some of the chemically modified viruses have 
different residual activities in various host plants. This may be due 
to differences in opportunity for penetration or in requirements for 


Several strains of bacteriophage have been highly purified. The 
purification process is relatively easy, since bacterial lysates contain no 
other elements of size comparable to the characteristic phage particles. 
Differential centrifugation has been mainly employed; precipitation 
methods have also been successful ( 499; 661 ) . Except for one unsub- 
stantiated claim for phosphorus-lacking phage (364), several coli- 
phages and one staphylococcus phage have proved to contain protein 
and DNA, the DNA in amounts as high as 40% or more of the dry 
weight. Amino acid analyses show small differences among related 
strains (table 10). The presence of lipids and RNA in some prepara- 
tions was originally reported (645), but careful work, including re- 
moval of traces of bacterial debris by serological precipitation (146), 
makes it likely that lipids and RNA are not essential constituents of 
virus particles. 

The structure and genetic complexity of bacteriophages indicate the 
great degree of differentiation that phage nucleoprotein must exhibit 
within each particle. Recent work has indicated that a large fraction, 
and possibly the totality of the protein of some large phages, forms an 


Chemical Composition of Vims Material 

. Table 10. Amino acid analyses of bacteriophages 

Analyses on preparations of bacteriophage purified by differential centrifu- 
gation. The degree of purity of all preparations is questionable. Note the rela- 
tively good agreement for preparations of a given phage analyzed by the same 

Preparations 1-5 analyzed by microbiological assay; Henderson, Sheek, and 
Luria (unpublished). Preparation P. and W. analyzed by paper chromatog- 
raphy; Poison and Wyckoff (&#>) All figures represent content of a given 
amino acid in the virus protein. 

Amino Acid 

Phage T2 

Phage T4 

Prep. 1 

Prep. 2 

Prep. 3 

Prep. 4 

Prep. 5 

P. and W. 

Aspartic acid 
Glutamic acid 



1 1 . 97 



























































outer coating that does not participate in phage reproduction (313; see 
chapter 8). If the nucleic acid should prove to be the true reproduc- 
tive material of these phages, we would have to visualize it as differ- 
entiated and organized in the complex way necessary to account for the 
genetic properties of the phages (see chapter 9). 

The dry weight of 1 infectious unit of a purified preparation of a 
large coli-phage is of the order of 10 ~ 15 gram, in good agreement with 
the weight calculated for the particles from their size. This, of course, 
reflects the high efficiency of the titration method (between 1 and 

CH. 5 Animal Viruses 105 

2.5 particles per infectious unit in good preparations; see page 52). 
The phosphorus content of a coli-phage such as T2 is approximately 
5 X 10 5 atoms per infectious unit, corresponding to 2 X 10 5 to 5 X 10 r> 
nucleotides per particle (319). The DNA of phages T2, T4, and T6 
of E. coli contain an unusual pyrimidine, hydroxymethyl-cytosine, in- 
stead of cytosine (6876). The new pyrimidine has not been isolated 
from any other source and may represent a unique constituent of these 
phages. Phage T2 contains 32.0 moles of adenine, 33.2 of thymine, 
17.9 of guanine, and 17.0 of the new pyrimidine per 100 moles of 

The discovery of hydroxymethyl-cytosine represents a remarkable 
advance not only in our knowledge of nucleic acids but also in our 
realization of the degree of specificity of the chemical processes that 
must be involved in virus synthesis and of the level at which an in- 
fecting virus must act upon the synthetic machinery of the host cell 
(see chapter 8). 


The purity of even the best preparations of animal viruses is open to 
question, because of the intrinsic difficulties involved in effective puri- 
fication and in the estimation of purity itself. On the one hand, the 
viruses must be extracted from tissues, in which they represent a small 
minority component. Normal components, in the form of particles of 
size similar to that of viruses ( microsomes ) , are also present and often 
contain nucleoproteins. The titration methods are seldom sensitive 
enough to permit any close correlation of one infective unit with an 
amount of virus of the order of one virus particle. Yet, analyses of 
several viruses carried out in different laboratories with different tech- 
niques have given reasonably consistent results, so that some generali- 
zations are possible. Most of the chemical information available con- 
cerns the largest viruses, purifiable either by centrifugation or by pre- 
cipitation (56). The smallest viruses can now be somewhat purified 
by removal of the larger impurities by protamine sulfate precipitation 
combined with trypsin treatment (664). 

No animal virus has yet been obtained in crystalline form, although 
viruslike particles extracted from certain forms of human warts show 
crystallike formations in electron micrographs (633). 

The chemical composition of some of the best preparations of several 
viruses is given in table 11. Opinions as to the validity of these 
analyses vary rather widely. It seems safe to assume that in the case 
of large, easily purifiable viruses such as vaccinia the chemical analysis 


Chemical Composition of Virus Material 


w -a 

1 I 

1 1 

S fe 


0^ C^ 

a I 

1 1 


<5 # 

cv. cv. <v. 


1 I 

^ ^ ^2 + 


00 W 00 "f O 
Q< l> t* OS O 



S S S 8 


O O O ^ ** CO 

Sod oo i>^ cd i>" 
OS O > t- t- 


i s 

(V co i t <v b- 


f* 00 O fr"~ t^ 
f-I CO t> GO O 


: 5 : =s 


S2 3 2 2* 

Whole Complex 

i J 

OO O *0 F-H O 

<^ co* * <v oo d 


O OS OS OS 00 

do ^ d do 


GO O *> O l> O 

d d > d os os 


GO ^< CO *O *O *O 

i j, g 

|l^fe e | 

*S o g ^ w g 'JJ 

CH. 5 Animal Viruses 107 

is fairly descriptive of the virus. DNA is probably the only constitutive 
nucleic acid in vaccinia virus. The evidence that some animal viruses, 
including those of the equine encephalomyelitis group, contain only 
RNA and no DNA is not fully convincing; nor is the evidence for the 
high lipid content of these viruses. Rous sarcoma virus appears to 
contain RNA (137). 

Little use can be made of the results of chemical fractionation into 
major components (proteins, nucleic acid, lipids, carbohydrates) for 
an understanding of virus properties. It seems possible that the infec- 
tious particles of most viruses will be found to contain only one type 
of nucleic acid ( either DNA or RNA ) , although excellent preparations 
of viruses of the influenza group yield both types (378). These highly 
purified preparations of influenza virus also contain some host-specific 
antigens, which might represent impurities, but which might also be 
incorporated into the particles as a necessary consequence of their 
mode of reproduction in the host cells (142; 377). 

The presence of DNA or RNA in the particles of a virus may suggest 
its possible natural relationship with nuclear or cytoplasmic nucleo- 
protein granules. The implications of this line of thought will be 
developed more fully in chapter 18. 

Amino acid analyses of repeatedly purified influenza virus prepara- 
tions (379; table 12) have revealed significant quantitative differences 
between virus types. These differences are more marked than those 
observed among different strains of tobacco mosaic virus. 

Even the purest preparations of some of the largest animal viruses 
show a complex composition resembling that of bacterial cells. Ele- 
mentary bodies of vaccinia virus, obtainable in amounts of 2 mg per 
rabbit, contain several protein fractions that can be separated and 
shown to be different by chemical and serological tests (606). The 
difficulty in deciding what really belongs to the virus is well illustrated 
by the presence, in purified vaccinia virus, of cholesterol in definite 
and constant amounts, which, however, can be removed by proper 
solvents without loss of activity. Should the cholesterol be considered 
an impurity? Or is its presence an indication of a role it plays in virus 
production? These questions will become more meaningful when bio- 
chemical studies of virus growth are more advanced. 

Other interesting questions are raised by the presence in purified 
vaccinia virus preparations of certain enzymes or coenzymes. Phos- 
phatase, catalase, and lipase activity are present in most virus prepara- 
tions. They may be due to enzymes accidentally adsorbed onto the 
virus particles during purification, since purified particles adsorb such 


Chemical Composition of Virus Material 

Table 12. Amino acid content of highly purified PR8 and Lee in- 
fluenza virus particles and of the sedimentable particles 
of normal allantoic fluid 

From Knight (379) 

Amino Acid 


(per cent) 


(per cent) 

(per cent) 


(per cent) 





Aspartic acid 
Glutamic acid 


















5 .5 





























. 05 

* Mean deviation of the values of single determinations from the average.*- 
given in the table. 

enzymes from solutions. Most careful tests for enzymatic activities of 
pure virus particles on simple substrates have been negative (41; table 
13). Adenosine triphosphatase has been reported in a fowl leucosis 
virus (486a). 

Biotin and riboflavine are present in constant and definite propor- 
tions in vaccinia virus and are apparently essential components of the 
virus particles (which also contain small but appreciable amounts oi 
copper: 0.05$). The role of these substances as coenzymes in some 
of the energy-yielding systems of many cells is well known. Theii 
presence in virus particles has been interpreted in the light of the 
theory that viruses may have originated from free-living forms. Sup- 
posedly, the viruses became obligate parasites through loss of certain 

CH. 5 

Animal Viruses 


Table 13. The occurrence of enzymes and other metabolic factors in 
the elementary bodies of viruses 

Modified. from Bauer (41) 






Adenosine nucleosidase 
Triosephosphate dehy- 


Probably absorbed onto virus 
from tissue of origin 

Cytochrome oxidase 
Cytochrome C 


Probably an integral part of the 
virus body 



Action on cell receptors 

Xanthme oxidaso 
Adonosine tnphospliatase 
Succinic dehydrogenase 

essential functions, for which they became dependent on the host. 
Thus, they may still occasionally possess some of the constituents of 
enzyme systems functional in their ancestors. Of course, the sub- 
stances in question might instead represent components of host-cell 
systems that are operative in virus synthesis, and that become incor- 
porated into the virus because of the peculiarities of its (unknown) 
mode of multiplication. 

As a whole, the absence of enzymatic activity on substrates of 
metabolic significance is in line with the view that viruses utilize, for 
their growth and reproduction, the energy-yielding machinery and 
the building-block-synthesizing systems of their host cells. Enzymatic 
activities to be expected in virus particles, if any, are more likely to 
be exerted on substrates that may be encountered by the virus in pene- 
trating or coming out of the host cell. In fact, the only well-established 

110 Chemical Composition of Virus Material 

enzymatic activity of virus particles is the receptor-destroying activity 
of the viruses of the influenza group and of other hemagglutinating 
viruses, which can destroy the surface receptors of red blood cells and 
other related substrates (330). The details of this enzymatic reaction 
and its possible role in virus-host cell relation will be discussed in 
chapter 13. 

A chemically well-investigated group of animal viruses is that of the 
insect viruses, which can be obtained in large amounts and consist of 
characteristic, easily purifiable rod-shaped particles (66). They con- 
tain DNA, protein, and little if anything else. In the specific inclu- 
sions, polyhedra, or capsules, the virus particles are associated with 
large amounts of a specific, noninfectious protein. This polyhedral 
protein consists of elements much smaller than virus particles, with a 
molecular weight of 200,000-400,000 (68a). The polyhedra, which 
are surrounded by a membrane, may be formed by precipitation of the 
specific protein, which is quite insoluble at its isoelectric point. 

Some types of polyhedra liberate no rod-shaped virus particles, give 
a pitted appearance after treatment with sodium carbonate, and ap- 
pear to have no membrane surrounding them (611). 

These findings emphasize the importance of investigating not only 
the infectious virus particles but also other "specific" substances found 
in virus-infected hosts in order to gain an understanding of the various 
phases of virus production and function. 


The evidence for such substances is rapidly increasing. For plant 
viruses, we have, for example, the noninfectious crystallizable protein 
found in plants infected with the Rothamstead strain of tobacco 
necrosis virus. This specific protein consists of homogeneous particles 
about one-half the diameter of the infectious spherical particles (49). 
Turnip yellow mosaic virus preparations (471; 473) can be separated 
into two fractions, both consisting of spherical particles of equal size, 
some (about 60S?) infectious and containing RNA, the others (about 
40%) without it. There are reasons to believe that the RNA-less par- 
ticles are not artifacts, but are actually present in the sap. They are 
antigenically similar to the virus, though less active. 

Similarly, bacteriophages and animal viruses are often accompanied 
by specific protein materials, which are often smaller than the viruses. 
Besides the polyhedral protein, we may list the soluble antigen of 
vaccinia virus (606), the complement-fixing soluble antigens of in- 

CH. 5 Specific, Noninfectious Substances 111 

fluenza and of several other viruses (383), and the "ultrafiltrate factor" 
in phage ly sates that blocks phage-neutralizing antibodies (111)- In 
the few cases in which chemical analysis has been possible, these sub- 
stances have proved to be proteins related to some of the proteins of 
the virus. Evidence is growing that at least some of these proteins 
represent intermediate products of virus synthesis; some may represent 
"mistakes" in virus production or breakdown products. Some such 
incomplete particles are sometimes produced instead of complete virus 
under the influence of chemical inhibitors. For example, bacteria 
infected with certain phages in the presence of the acridine dye pro- 
flavine liberate incomplete, nucleic acid-free particles (175). 

There may be a whole range of cases intermediate between the pro- 
duction, along with the virus, of an excess of small building blocks and 
the production of a mixture of active and inactive virus particles. It 
is also possible that some of the specific, noninfectious virus products 
derive from the breakdown of virus particles. The amounts of these 
products vary greatly. For bacteriophages, the "ultrafiltrate factor" 
represents a small fraction of the total virus material produced; the 
virus appears to waste little in byproducts. In the polyhedroses of in- 
sects, on the other hand, the virus is a small fraction of the specific 
material of the polyhedra, the noninfectious bulk of which may fulfill 
some protective function for the virus inside the cell. 

These examples emphasize the importance of not limiting our atten- 
tion to the infectious particles themselves. Even the fact that virus 
particles may be inactive, when extracted from their host cells, because 
they are combined with inhibitors of cellular origin, may suggest a 
possible role of these inhibitors in virus production. 

The fact that at least some of the "specific soluble substances" of 
viruses are proteins without nucleic acid and can be demonstrated 
separately from the virus particles raises a point of some interest. Are 
the same proteins, when in the virus particles, bound to nucleic acid 
or not? The ease of extraction of soluble antigen from vaccinia virus, 
as compared with the difficulty of removing any nucleic acid, seems to 
answer in the negative. Thus, the proteins of a virus particle may con- 
sist partly of nucleoproteins and partly of other proteins without 
nucleic acid ties. 

As a whole, chemical investigations on purified viruses have as yet 
shed little light on the basic problems of virology. The fault rests 
neither with the chemists who have worked on these problems nor 
with any peculiarities of viruses, but with the disparity between the 
questions asked and the techniques available. The goal is to interpret 

112 Chemical Composition of Virus Material 

biological functions, such as reproduction, host-range specificity, and 
tissue specificity, in terms of chemical structure. Present-day chem- 
istry, however, lacks the proper tools to supply the answer. Protein 
and nucleic acid structure is understood only in some of its grossest 
phases, and nothing is known of the chemical basis of those specific 
properties of a protein that are of biological importance, such as anti- 
genicity, species specificity, and enzyme function. The present gap 
in our chemical knowledge will be filled only by the development of 
a new biochemistry, covering the area of specific differentiations of 
large organic molecules of biological importance. 


The crystallization of plant virus proteins started widespread specu- 
lations and controversies as to the nature of viruses, some workers 
considering them "molecules," and therefore "nonliving," and others 
considering them "living organisms." The emotional root of the con- 
troversy lies in the reluctance of the chemist to call living something 
simple enough to crystallize in a test tube, and on the hesitance of the 
biologist to admit the existence of molecules capable of reproduction. 
The semantic aspect of the controversy, however, is more important, 
since it illustrates the fact that both the terms "molecule" and "living" 
must be redefined before we can utilize them as categories in which 
entities such as viruses can be classified unequivocally. 

On the one hand, no definitions of the word molecule, as given by 
the chemist, can be applied without ambiguity to particles of bio- 
logical material. Actually, no protein has ever been synthesized, the 
details of protein structure are obscure, and the structural homogeneity 
of even the most purified protein is questionable. Like viruses and 
visible cell components, proteins are synthesized only inside living 
cells. The term "molecule" is inadequate as a pigeonholing device in 
classifying such objects. 

On the other hand, the confusion that would arise if all proteins 
were classified as "living" in view of their cellular origin is illustrated 
by a simple example : could we consider a preparation of egg albumin 
as "living"? This suggests that the limits of applicability of the term 
"living" in its everyday, common-sense meaning become hazy as we 
reach the domain of protein chemistry. 

We should, then, investigate the possibility of redefining the word 
"living" in a way that may be meaningful and useful in this borderline 

CH. 5 Living Nature of Viruses 113 

field. The redefinition must aim at selecting criteria for the classifi- 
cation of objects in such a way that fewer and less important am- 
biguities will arise. As an analogy, the reader may consider the prob- 
lem of adopting criteria that make the distinction between "table" and 
"chair" as useful as possible even in the borderline field of flat-top 

A point of departure may be the consideration that all properties of 
biological systems reflect their material structure and are not mani- 
festations of metaphysical "life forces." Life, however defined, will be 
a manifestation of certain organized material objects that we call "living 

Throughout the physical world, we observe that certain new prop- 
erties arise as a result of organized agglomerations of parts, or, as is 
usually said, at certain levels of integration, not because the new 
properties are metaphysically superimposed over and beyond the prop- 
erties of the component parts, but because the observations by which 
they can be defined can be made only on the whole and not on the 
isolated parts. For example, the social behavior of an individual can- 
not be defined in terms of the individual taken in isolation, outside of 
any social situation. The requirement for operational definitions, that 
is, for criteria based on feasible observations, can also be illustrated 
with the well-known example of the clock. All mechanical properties 
of a clock reflect the mechanical properties of its component parts, but 
the meaning and the very existence of the clock as a time-measuring 
device only arise upon its being assembled and observed by a time- 
minded agency. 

Having decided that life must represent a manifestation of certain 
organized portions of matter (organisms), we realize, however, that 
each individual property of organisms, as listed in textbooks of biology, 
becomes somewhat ambiguous at the borderline between simple or- 
ganisms and large units or molecules of organic matter. Some writers 
have, therefore, despaired of the feasibility of defining life. Rejecting 
such pessimism, we may consider that a property of all "unambiguous 
organisms" is their ability to reproduce, that is, to introduce a struc- 
tural and functionally specific organization similar to their own into 
other material of simpler chemical organization. 

The property of reproduction, however, if used as the sole criterion 
for defining life, would face two objections. First, crystal growth some- 
times results in the formation of chemical complexes that do not exist 
as such in the mother solution, and, therefore, in an increased degree 
of specific chemical organization. Second, the nature of the synthetic 

114 Chemical Composition of Virus Material 

mechanisms underlying biological reproduction is unknown, and it may 
be suggested, for example, that some or all proteins, in their native 
intracellular state, take part in reproduction. The simple criterion of 
reproduction is, therefore, of little value in formalizing the differences 
between groups of objects in the field under discussion; it fails because 
it is not strictly definable in terms of observable events and per- 
formable operations. 

A further approach can be made, however, by taking into considera- 
tion the historical aspect of biological phenomena. There exists in the 
material substrate of the biological world a unity derived from its 
common origin and continuity. This unity is embodied in the two 
fundamental tenets of biology: the rejection of spontaneous generation 
and the theory of evolution. Unity and continuity reflect the presence 
of certain specific materials, which regulate the assimilation of outside 
matter and its transformation into more of the specific components. 
These specific materials, capable of change and evolution, represent 
the elements of continuity in the cycle of organic matter, by which 
simpler, nonspecific building blocks are organized into more proto- 
plasm. A satisfactory operational definition of a living substance or 
organism might, therefore, be the following: "A material is living if, 
after isolation, it retains a specific configuration that can be reinte : 
grated into thet cycle of organic matter." 

This definition, by making life identical with the possession of an 
independent, specific, self-replicating pattern of organization, gives a 
tangible criterion of classification. It differentiates, for example, be- 
tween most cell proteins, which once isolated are unable to reenter the 
cycle of organic matter without being broken down below the level of 
biological specificity, and the viruses, which can reintroduce their own 
pattern into the network of biological syntheses. It also excludes, 
deliberately, the individual cells of multicellular organisms unless they 
are capable of reproducing more cells, for example, in tissue cultures. 

Our definition does not involve any assumption as to the mechanism 
by which the pattern of specificity is preserved and reproduced or as 
to the influences that may alter it. It requires only that any change 
or mutation that maintains the living character must give rise to a 
new self-replicating pattern. The operational character of the defi- 
nition is illustrated by the fact that a virus, for example, can only be 
considered as "more living" than any other cell constituent extracted 
along with it by performing the proper test of reintroduction into a 
suitable environment. 

CH. 5 Living Nature of Viruses 115 

It will be realized that our definition classifies as living such entities 
as the "transforming principles" responsible for induced specific bac- 
terial changes (210), but excludes, among other things, individual 
genes that cannot (yet?) be isolated and reintegrated into cells (see, 
however, page 351; 32; 695). 

It should be pointed out that some of the areas left ambiguous by 
our definition involve such biological enigmas as those plant diseases 
supposedly caused by viruses but transmissible only by grafting. The 
only reason to suspect the presence of a virus is the repetition of an 
abnormal symptomatology. Whether such hypothetical "viruses" that 
cannot be isolated are to be called living is anybody's choice. 



Serological Properties of Viruses 

Virus particles contain large amounts of protein. Since most pro- 
teins are good antigens, it is to be expected that viruses will have 
antigenic properties. An antigen is defined as a substance that, intro- 
duced parenterally (not by mouth) into an animal organism, stimulates 
the production of antibodies, that is, of modified serum globulins which 
circulate in the blood serum and can react specifically with the an- 
tigen. An antigen-antibody reaction involves combination between 
the two, in the case of viruses as well as of other antigens. Serological 
reactions are specific. Their specificity, determined by the chemical 
nature of the antigen, is of very high grade, since it distinguishes 
among protefns whose chemical differences are well beyond the ana- 
lytical methods of present-day chemistry. Serological reactions, there- 
fore, can be considered as a refined tool for the structural analysis of 
proteins and other antigens and of particles containing such antigens. 

Reactions between an antigen and an antiserum prepared against 
another antigen (heterologous as opposed to homologous reactions) 
indicate the possession by the two antigens of common or very similar 
chemical configurations. Incomplete antigens or haptens are sub- 
stances that, although incapable of stimulating antibody formation, 
react with antibodies against antigens with which they share certain 
chemical configurations. 

Since all the specific properties of a virus, whether in the extra- 
cellular or in the intracellular state, ultimately depend on the chemical 
structure and organization of the virus material, Serological methods 
are a precious tool for the study of viruses. We have already men- 
tioned that the present methodology of protein chemistry and physico- 
chemistry including elementary and amino acid analysis, sedimenta- 
tion, diffusion, light-absorption measurements, and electrophoresis 
tells us very little about the structural basis of the biological specificity 


CH. 6 Antigen- Antibody Reactions 117 

of proteins. This specificity must depend on details of organization, 
and folding of polypeptide chains, details which disappear in the 
course of chemical analysis and which are not revealed by physico- 
chemical methods. Serological analysis, however, represents a finer 
chemical tool. An antibody is a chemical reagent capable of distin- 
guishing those details of organization on which functional specificity 
may depend. Indeed, antibodies may distinguish between proteins 
prepared in identical ways from the same tissue of different related 
species or even of different individuals of the same species (397). 

Each virus contains specific antigens, distinct from those of the 
host in which it grows and from those of other viruses. Antiviral 
antibodies permit not only the recognition and identification of indi- 
vidual viruses but also the recognition and measurement of virus 
material inside the host, even when this material is not extractable in 
an infectious state. Serological methods allow us to recognize simi- 
larities and differences among virus strains. They provide us with 
useful criteria for virus classification and with valuable epidemiological 
information. Serological reactions are the basis for most of the diag- 
nostic procedures in virus diseases of man and animals. The animals 
used for laboratory production of antiviral antisera are generally rab- 
bits, although guinea pigs, chickens, ferrets, and horses are sometimes 
used for special purposes. 


When an antigen is mixed with homologous antiserum, the reactions 
observed depend on the physical properties of the antigen and on the 
multiplicity of distinct antigenic functions it possesses. The types of 
reactions by which antigen-antibody combinations are manifested may 
be divided into three groups. First, there are reactions in which we 
observe changes in the state of dispersion of the antigen, such as pre- 
cipitation or agglutination. In the precipitate, individual molecules 
or particles of the antigen are held together by antibody molecules in 
a framework or "lattice." According to widely accepted immuno- 
chemical theories, both antigen and antibody must be multivalent or 
at least bivalent to allow formation of such a lattice, since each antigen 
particle must combine with at least two molecules of antibody, and 
vice versa, to bring about formation of an extensive framework. The 
extent of lattice formation depends on the relative concentrations of 
the reagents. There will be an optimum ratio for lattice formation 

118 Serological Properties of Viruses 

(equivalence point), and a lattice may fail to be formed if either 
reagent is in large excess. The term precipitation is used for antigens 
such as protein molecules; agglutination refers to larger particles such 

10 16 

10 14 

1 10 12 

x Diody 0.001 mg 

io l 
io 8 

IO 4 
IO 2 

Experimental points 

a Theoretical points 
Experimental threshold serological reaction curve 

Theoretical threshold serological reaction curve on 

a basis of 0.001 mg antigen per cubic centimeter 

OPoliomyelitis virus 
^^Tobacco mosaic virus 


virus <,. 

fellow fever virus "* 

\Phage (Burnet) 
xS^Borna disease virus 
Phage (Burnet)* ^ \SiVaccinia 

Vaccinia ^Bovine pleuropneumoma organism 

'**_ \ * Paratyphoid bacilli 

\ \lo.01 mg 
>, \ 

* \ 

Oh. v 

\ \ Red bloodV 

* \ cells 0.1 m> 


Red blood cells \ 
0.001 mg 

id" 18 

id" 16 

I I I I I I 








id" 2 

id" 10 

l<f 8 


Mass of each particle in milligrams 

Figure 49. The relation between the mass of a particle and the number of par- 
ticles needed for visible precipitation or agglutination by specific antiserum. From: 
Merrill (480). The mass values for several viruses should be somewhat modified 
in the light of present-day information. 

as bacteria. The distinction is rather immaterial in the case of antigens 
such as virus particles. Large amounts of antigen are needed for the 
production of visible reactions; with viruses, the minimum amount 
required is of the order of 0.1 to 1 /ug of virus material (480). Assum- 
ing the need for a constant minimum mass of antigen, a relation can be 

CH. 6 Antigen- Antibody Reactions 119 

derived between the size of virus particles and the number of particles 
required for visible precipitation, as shown in figure 49. 

Agglutination of virus particles may be visualized microscopically 
or ultramicroscopically, for example, in the electron microscope. A 
modified test is the agglutination by antiviral serum of relatively large 
particles, such as collodion granules or bacterial cells, which have been 
artificially coated with virus. The coating may be aspecific, as for 
vaccinia virus on collodion particles (258), or specific, as for bacterio- 
phage on heat-killed bacteria (121). 

The second type of reaction is one in which, after mixing antigen 
and antiserum, we test for a modification of some properties of the 
serum due to combination of the antibody with the antigen. Typical 
of this is the complement-fixation reaction, based on the fact that, in 
combining together, antigen and antibody often bind a normal con- 
stituent of blood serum, called complement. A complement-fixation 
test is made by adding to the mixture of antigen and antibody a known 
amount of complement, that is, of normal guinea pig serum. The dis- 
appearance or fixation of complement is then tested by means of an 
"indicator system," which consists of sheep red blood cells treated with 
antisheep cell serum. If complement is present, hemolysis of the cells 
occurs; absence or reduction of lysis shows that complement has been 
removed. The complement-fixation test is of great diagnostic value in 
animal virus diseases, in testing sera of suspected patients for the pres- 
ence of virus antibodies. The virus preparations used as antigens in 
the complement-fixation test need not be as pure and concentrated as 
for agglutination or precipitation tests. Complement fixation, how- 
ever, does not occur in all antigen-antibody reactions. 

The third type of reaction involves changes in the infectivity of the 
virus as a result of combination with antibody. This is the so-called 
neutralization reaction, in which we observe a reduction in the infec- 
tious titer of virus. Loss of virus activity as a result of combination 
with antibody suggests that some essential portion of the virus particle 
is prevented from functioning by combination with antibody. This 
indicates the possibility of identifying specialized structures in the 
virus with certain antigens. As already mentioned, an antigenic func- 
tion depends on the presence in the antigen of certain specific chemical 
groups or configurations, the "antigenic determinants," which induce 
the formation of antibody molecules capable of specific combination 
with these determinants. These may vary in complexity from rather 
simple functions, such as aromatic groups, to large portions of a pro- 

120 Serological Properties of Viruses 

tein molecule. Complex entities such as virus particles i lay carry 
several distinct antigenic determinants. An antiserum prepare*! against 
a virus may then contain a variety of specific antibodies. 1 

The multiplicity of viral antigens can be illustrated for vaccinia virus 
(606). On the one hand, injection of intact virus particles gives anti- 
sera that neutralize and agglutinate virus particles. On the other hand, 
antigenic fractions can be obtained which give rise to antibodies that 
agglutinate but do not neutralize the virus particles. It is clear, then, 
that this virus contains at least two types of antigens. The tests for 
multiple antigens are often delicate. For example, precipitation of a 
specific antiserum with particles of papilloma virus removes all anti- 
body activity from the serum (105). The same is true for bacterio- 
phage particles and antiphage serum. Yet, we can obtain phage 
antigen, in the form of immature particles, that react with antiphage 
antiserum but do not combine with phage-neutralizing antibody (175). 
The most probable interpretation of virus neutralization is that 
combination of a virus particle with antibody blocks certain surface 
groups that are required to initiate infection. Combination with anti- 
body apparently does not destroy the virus. It is often possible to re- 
cover active virus upon removal of antibody by a variety of treatments. 
Several treatments suppress infectivity without altering the anti- 
genicity of viruses. This indicates either that the changes involved in 
inactivation by these treatments are of a finer chemical nature than 
can be detected by tests of serological specificity, or that the groups 
affected in the inactivation process are not directly involved in the 
determination of antigenic specificity, that is, are not part of antigenic 
determinants. This is understandable, since inactivation may result 
from slight changes in one localized point of a virus particle. Usually 
the treatments that leave the virus antigenic but not infectious (for 
example, formalin or radiations) are those that inactivate the virus 
without extensive protein denaturation. 

The problem of the possession of multiple and different antigenic 
determinants by certain viruses is connected with the problem of cross- 
reactions. It is generally found that a serum reacts more strongly or 
at higher dilution with its homologous antigen than with heterologous 

1 This fact is not necessarily reflected in the differences of the reactions observed, 
since the reactions depend entirely on our methodology. For example, there is no 
a priori reason to consider the agglutinating and the neutralizing antibodies as 
different antibodies, since agglutination and neutralization of virus particles may 
both be due to combination of the same antigenic determinant with the same anti- 
body, the combination being tested in different ways. 

CH. 6 Antigen-Antibody Reactions 121 

antigens. If any cross-reaction occurs at all, the difference between 
homologous and heterologous antigens is clearly proved by absorption 
tests. Absorption with the homologous antigen generally removes 
from a serum all antibodies, while absorption with a heterologous 
antigen may only remove part of the antibodies, leaving a certain 
amount of reactivity with other antigens. The situation is illustrated 
for a group of related plant viruses in table 14. A virus could, thus, 

Table 14. Cross-precipitation experiments with plant viruses and 
cross-absorbed antisera 

From Bawden (43). + , + + , 

+ + + : degree of precipitation at optimum antigen-anti- 
serum ratio. 

Precipitation Tests 


Used for 
















Virus 3 

Virus 4 

Tobacco mosaic virus 









+ + 

+ 4 



+ + + + 

4 + 44 

+ + + + 



+ + + + 

4 + 44 

+ + + + 



Aucuba mosaic virus 



+ 4 

+ + 





+ + 






+ 4-4- + 

444 + 

+ + + + 




+ + + + 

4 + + + 

+ + + + 



Enation mosaic virus 




+ + 







+ + + + 

+ + + + 




+ + + + 

+ + + + 



Cucumber mosaic 





+ + + + 

+ + + + 

virus 3 


+ + + + 

+ + + + 


+ + + + 

+ + + + 







be considered a mosaic of several discrete antigens, some shared and 
some unshared by related strains. 

This formal interpretation, however, may be misleading. Cross- 
reactions of this type are not always the expression of a mosaic of 

122 Serological Properties of Viruses 

discrete antigens, but may represent the fact that the virus strains 
possess antigens similar but not identical to each other. The anti- 
bodies produced as a response to a given antigen may consist of a 
variety of antibody molecules with varying affinity for the common 
antigen. Heterologous antigens remove only those antibody molecules 
that have a stronger, less discriminating affinity, leaving the weaker, 
more specific ones behind. Cross-reactions may then be considered 
as an indication of chemical similarity and relation, but not neces- 
sarily as proof of the possession of a number of identical antigenic 
groups in common. 


Bacteriophages illustrate well the serological specificity of viruses, 
since practically every strain of bacteriophage isolated in nature is 
found to be an antigenically distinct entity. Some bacteriophages 
are better antigens than others, as measured by the potency of the 
antisera that can be obtained against them. Two bacteriophages 
active on the same bacterium are often completely distinct serologi- 
cally. Cross-reactions generally occur only among bacteriophages that 
are similar in other properties, for example, in morphology, and the 
strength of the cross-reactions may be taken as an indication of the 
degree of relatedness. As a consequence, serological properties pro- 
vide a convenient basis for phage classification. For example, a large 
number of coli-dysentery phages fall into eleven major serological 
groups that do not cross-react (110). No serological relation exists 
between phage and host; antibody against the host gives no evidence 
of combination with the phage, and vice versa. A typical group of 
related phages includes the T-even (T2, T4, T6) series of coli-phages 
(169). Results of cross-neutralization experiments are given in 
table 15. 

Agglutination reactions can be observed with highly concentrated 
phage suspensions, particularly if the phage is a "large-particle" one, 
or with phage-coated bacteria (121). The neutralization reaction 
is, however, far more important, particularly because the titration 
methods for phage allow precise quantitative studies of the neutrali- 
zation process. 

The basic finding is expressed by the so-called "percentage law," 
which states that a given concentration of antiserum in a given time 
will inactivate a given proportion of the phage present, independent 
of the total amount of phage (29). This law holds true as long as 

CH. 6 Serology of Bacteriophages 123 

the amount of phage present is small as compared with the amount of 
antibody, so that the antibody available to each particle is not in- 
fluenced by the number of particles present. 

The second rule is that phage inactivation is a practically irreversible 
reaction. Serum-inactivated phage is not reactivated by dilution. 
Some phages, however, can be reactivated by sonic vibration (21) or 
by treatment with papain, which digests the antibody (366). 

The third rule is that the logarithm of the fraction of phage that 
remains active after a certain time in a mixture with antibody is pro- 
portional to the time of contact and to the concentration of antibody. 
This is expressed in the following equations: 

p _ A) P 

= e ktc \og e - = ktC Iog 10 = QASktC [9] 

I o P P 

where P is the initial phage activity; P is the residual active phage 
after t minutes of contact with a concentration C of antiserum; k is a 
constant ("fractional rate of inactivation"), which, within the limits of 
validity of equations 9, characterizes the rate of inactivation of a given 
phage by a given serum. Experimental data illustrating this relation 
are given in figure 50. 

The interpretation of the relation is not completely clear. The equa- 
tion is that of a first-order reaction or of a bimolecular reaction with 
one reagent in large excess. It suggests that one antibody molecule 
can bring about inactivation of a virus particle. If several molecules 
were needed, there would be an initial lag in the rate of phage in- 
activation, when few or none of the particles have as yet combined 
with the minimal numbers of antibody molecules necessary for inacti- 
vation. Not every molecule of antibody that combines with phage 
inactivates it, however. The phage particles that survive antibody 
treatment are already modified by combination with antibody. They 
are adsorbed less rapidly by sensitive bacteria and sometimes are not 
adsorbed at all by some of their host bacterial strains (365). Upon 
ultrafiltration, they show an increased size. Probably, antibody mole- 
cules can combine with different sites of the phage surface, but only 
the molecule that happens to block a certain spot suppresses the 
activity. This spot may be unique on the phage particle, or it may 
become singled out a posteriori, in the sense that, if a phage utilizes 
for attachment to a bacterium a serum-inactivated spot, the result 
may be a failure to grow. Indeed, it is known that some serum- 
inactivated phages retain at least part of their ability to be adsorbed 
by bacteria. 

124 Serological Properties of Viruses 

Equation 9 is not always strictly followed. An initial lag in inacti- 
vation is often present for low serum concentrations. Another frequent 
deviation is the occurrence of a more resistant fraction of phage, which, 
however, is not genetically resistant, since it does not give rise to a 


20 40 60 


Figure 50. Neutralization of bacteriophages by homologous or hetcrologous 
antiserum. The fraction F/F of residual infectious phage as a function of time 
of exposure to antiserum at 37 C. : Phage T2 + anti-T2 serum diluted 1:64,000. 
o: Same, diluted 1:16,000. D: Same, diluted 1:4000. x: Phage T4 + anti-T2 
serum diluted 1:4000. 

resistant progeny. The resistant fraction may represent physiologically 
different phage particles. It might possibly consist of particles that 
have combined with antibody molecules without being inactivated and 
in such a way that some necessary portion of their surface, although 
still unblocked, has become protected from further antibody molecules 
by steric hindrance on the part of the antibody already adsorbed. 

Analysis of precipitates formed by phage and antibody shows that 
some of the large bacteriophages, for example, coli-phage T2, can take 
up as many as 90 antibody molecules before losing activity. Up to 

CH. 6 

Serology of Bacteriophages 


5000 antibody molecules of rabbit antiserum may be adsorbed by 1 
phage particle (317). Absorption of a serum with a sufficient amount 
of phage, active or inactivated by formalin, can remove all the neu- 
tralizing antibody of the serum. 

Equation 9 is of practical importance in phage research. Once the 
value of fc, the fractional rate of inactivation, is known for a given 
serum-phage combination, we can calculate the serum concentration 
that will give any desired amount of survival in the desired interval 
of time. The values of k for a given antiserum and different phages 
indicate the degree of serological relationship among the phages. An 
unrelated phage will give a k value of 0; a weakly cross-reacting phage 
will have a k value much lower than that of the homologous phage 
(table 15). 

Table 15. Cross-neutralization by representative antisera against 
bacteriophages T1-T7 

A* = fractional rate of inactivation, equation 9 (min *) 































Although serum-inactivated phage cannot be reactivated by dilution, 
it has been rendered active again by digesting away the antibody with 
the proteolytic enzyme papain (366). This indicates that the virus 
particle was coated by antibody but not destroyed. 

An important observation (which appears to be valid for all viruses) 
is that phage-neutralizing antibody cannot prevent the multiplication 
of phage that has already been taken up by sensitive cells (167). 
This fact will be discussed again in chapter 8. In line with this ob- 
servation, it is impossible to eliminate with antiserum the phage carried 
inside lysogenic bacteria, which continue therefore to carry latent phage 
after having grown in antiphage serum. 


Serological Properties of Viruses 

Complement fixation by phage-antiphage serum combinations oc- 
curs. Its specificity, as measured by the extent of cross-reactions, is 
similar to that of neutralization reactions with the same serum (399a). 
The role of complement in the neutralization reaction itself is obscure. 
Addition of complement reduces the serum-neutralization rate for 
phage Tl, for example, but increases the rate for phage T2 (311). 

Phage particles are not the only phage-specific antigen. Lysates of 
several phages contain a smaller, ultrafiltrable component, which com- 
bines specifically with antiphage serum even when the antiphage 
serum has been produced by injection of purified phage. These 
phage-specific soluble substances or ultrafiltrable factors are detected 
by their serum-blocking power, that is, by their ability to reduce the 
phage-neutralizing activity of a serum (see table 16). They have the 

Table 16. The serum-blocking power of phage and of ultrafiltrate of 

phage lysate 

From DeMars et al. (175) 

Preincubation of antiphage serum with ultraviolet-inactivated homologous phage or with the phage- 
free ultrafiltrate of a lysate of the homologous phage reduces the phage-neutralizing ability of the 



of Test 

Per Cent 



of Test 








7.7 X 10 6 



Serum * 


AQO (^ 




4R C* 

2.2 X 10 4 



Serum + 7.0 X 10 9 T2 (UV inactivated) 

7.7 X 10 6 



Serum + 2.3 X 10 9 T2 (UV inactivated) 

4.8 X 10 5 



Serum + 7.7 X 10 8 T2 (UV inactivated) 

1.8 X 10 5 



Serum + 2.6 X 10 8 T2 (UV inactivated) 

4.3 X 10 4 



Serum + 8.7 X 10 7 T2 (UV inactivated) 

1.6 X'lO 4 



Serum + ultrafiltrate diluted 1 : 3 

1.7 X 10 5 



Serum + ultrafiltrate diluted 1 : 9 

5.3 X 10 4 



Serum + ultrafiltrate diluted 1 :27 

2.5 X 10 4 


* Serum anti-T2 1 : 40,000. 

same serological cross-reactivity as the corresponding phages. They 
are of interest because they may represent specific phage building 
blocks and may throw some light on the process of phage synthesis. 
Immature phage particles produced by infected bacteria in the 
presence of the dye proflavine and resembling empty phage heads 

CH. 6 Serology of Plant Viruses 127 

(see page 71 and figure 28), fix complement with antiphage serum 
but do not combine with neutralizing antibodies (175). Thus, there 
is more than one type of antigen in some phages. The antigen that 
gives rise to neutralizing antibodies is apparently localized in the 
phage tail, whereas another antigen, recognizable by complement- 
fixation tests, is located in the phage head (399a). Antibody against 
the head antigen does not neutralize the phage. These observations 
are in line with the conclusion (see chapter 8) that the phage tail 
represents the organ of attachment of phage to the host cell 


The antigenic properties of plant viruses were discovered relatively 
recently (536). The serological methodology is influenced by the fact 
that relatively large amounts of virus material are often available, for 
example, for tobacco mosaic virus. It is feasible, therefore, to run 
precipitation tests. These can actually be used to measure the amount 
of virus material in a preparation (see 43). Two methods are em- 
ployed: In the optimum precipitation method, a constant amount of 
virus suspension is mixed with various concentrations of antibody, and 
the mixture in which precipitation first appears is considered as the 
optimum for the reaction. In the end-point method, one determines 
the smallest amount of antigen that reacts with a constant amount of 
antibody. In these tests, as in all precipitation reactions with plant 
viruses, it is important to distinguish between specific and aspecific 
precipitation, since the viruses tend to be precipitated out of solution 
by a number of foreign materials, particularly proteins. 

Antisera have been obtained against several plant viruses. Those 
viruses against which no antibodies have been found may simply have 
been present in amounts too small to give visible precipitation. Com- 
plement fixation, which has been successfully applied to plant virus 
work (641a), may be useful in detecting smaller amounts of virus. 

Neutralization of plant viruses by antibody must be distinguished 
from the aspecific neutralization which occurs with normal sera. It is 
generally recognized that specific neutralization occurs, but that the 
combination of virus with neutralizing antibody is somewhat reversible 
upon dilution of the mixture. Tobacco mosaic virus can be recov- 
ered in active form from a mixture with antibody by digesting the 
antibody with enzymes. Quantitative studies on the rate of neutrali- 
zation have been few and relatively inconclusive. They are compli- 
cated not only by the aspecific neutralization by normal serum, but 

128 Serological Properties of Viruses 

also by the fact that, due to the low efficiency of the methods for 
testing infectivity, large amounts of virus must be used, so that the 
concentrations of antigen and antibody are of the same order. There- 
fore, the results are very sensitive to small variations in the concen- 
tration of either reagent. 

Since viruses like tobacco mosaic virus are available in large amounts 
we can easily perform absorption tests, in which the antibodies that 
react with a virus are removed by prolonged treatment with an excess 
of that virus, and then the reaction of the antiserum with a different 
virus is tested. By this method, the serological cross-reactions among 
plant viruses have been studied in some detail. The patterns of rela- 
tionship obtained in this way are in agreement with those derived 
from other methods, especially cross-protection tests. The results of 
a typical cross-precipitation test with absorbed antisera are shown 
in table 14. 

Crude preparations of plant viruses, like those of bacteriophages and 
of animal viruses, often contain materials that share the serological 
specificity of the virus particles but not their infectivity (49; 473). 
These materials may even be antigenically more powerful than the 
virus itself. Their exact relation to the virus particles, as precursors, 
breakdown products, or abnormal forms, is currently a matter for 


The serological reactions of animal viruses, discovered very early 
after the recognition of virus diseases, comprise a major portion of 
medical virology. Antibodies play an important role in the course of 
animal virus diseases and in a variety of important procedures in the 
prophylaxis, diagnosis, and therapy of these diseases. Serological 
reactions have been studied both in animals that can support virus 
growth and in animals that cannot. In those that can, the role of 
antibodies in immunity phenomena has been thoroughly investigated. 

In vivo tests of the infectivity of mixtures of a virus with the 
homologous antiserum show a reduction of activity, commonly inter- 
preted as due to a combination of the virus with neutralizing anti- 
bodies. This interpretation had been challenged because the com- 
bination is difficult to demonstrate. Indeed, virus activity can often 
be recovered from neutralized mixtures with antibody either by dilu- 
tion or by centrifugation. It was, therefore, suggested that antiserum 
acts by stimulating the resistance of the host tissues (570). Good 
evidence has been obtained, however, for actual combination of the 

CH. 6 Serology of Animal Viruses 129 

virus with neutralizing antibodies (see 121). In the first place, the 
recovery of activity from virus-antiserum mixtures becomes progres- 
sively less complete the longer the mixtures are allowed to stand. 
Second, actual absorption of the neutralizing power of a serum by 
washed virus particles has been demonstrated. Burnet and his col- 
laborators, using for virus titration the accurate method of pock count- 
ing on the chorioallantoic membrane of the chick embryo, showed 
that the neutralization of a variety of viruses by antiserum can best be 
interpreted in terms of a reversible combination between virus and 
antibody. The reversibility is more or less complete depending on the 
virus. With rabbit myxoma virus, for example, concentrated antisera 
give irreversible neutralization, whereas with louping ill virus the 
reaction seemed to be completely reversible. With influenza viruses, 
it can be shown by absorption tests that the neutralizing antibodies 
are at least in part the same as those that prevent red blood cell ag- 
glutination by the virus in vitro, under conditions where the antibody 
must certainly act on the virus itself. 

It is now certain that the mechanism of action of virus-neutralizing 
antibodies is not fundamentally different from any other antigen- 
antibody reaction. The outcome of an inoculation with a virus-anti- 
serum mixture depends upon the availability of the surface groups of 
the virus particle needed for infection at the time of the contact of 
the virus with a susceptible cell. This conclusion is supported by 
the finding that intradermal injection of antiserum against vaccinia 
protects a rabbit against dermal infection if given 24 hours previous 
to, or simultaneously with, the injection of virus, but not if given as 
little as 5 minutes after virus injection at the same site. This illus- 
trates the inability of antibodies to affect virus that has already been 
taken up by the host cells. In general, the outcome of inoculations 
with mixtures of virus and antiserum will depend on the net balance 
of several processes: the reversibility of the virus-antibody combina- 
tion, the rate of uptake of active virus by sensitive cells, and the rates 
of spread of the virus, the antibody, and the virus-antibody complexes 
through the tissues. 

A possible biological significance of the reversibility of the combina- 
tion of animal viruses with neutralizing antibodies has been suggested 
by Burnet (121). Reversibility favors the survival of the virus, both 
in its spreading within the host organism and in passing from host to 
host. The ability to combine reversibly with antibodies, being advan- 
tageous to animal viruses, would be favored by natural selection, 
whereas for bacteriophages irreversibility of the combination with 

130 Serological Properties of Viruses 

antibody would not be a handicap for survival in nature, where phage 
does not generally meet antibody. 

In vitro serological reactions of animal viruses, such as agglutination, 
precipitation, and complement fixation, reveal the antigenic complexity 
of most viruses. Crude virus preparations often contain, besides the 
virus particles, other virus-specific antigens. A good example of these 
noninfectious virus antigens, often called specific soluble substances 
(SSS), is the SSS of vaccinia virus (606). Centrifugation of a crude 
preparation of vaccinia from the rabbit's skin leaves a supernatant 
which, after being freed of residual elementary bodies, contains large 
amounts of a specific protein antigen, the so-called LS antigen. After 
purification, the LS fraction gives precipitation and complement- 
fixation reactions with antisera prepared either against the virus 
particles or against pure LS. The LS antigen has been found to 
consist of protein molecules, about 300,000 in molecular weight, carry- 
ing two distinct and separable antigenic functions L and S. The L 
function is suppressed by heating, the S function by treatment with 
chymotrypsin, The two antigenic functions are carried on the same 
molecule. This is shown by the fact that complete LS antigen is fully 
precipitated by antisera prepared against each of the partial antigens, 
L or S. Antibody against LS also agglutinates elementary bodies, indi- 
cating that some LS antigen is probably located on the surface of the 
virus particles, from which it can actually be extracted. The antibody 
against LS does not, however, neutralize the virus. The LS antigens 
of all viruses of the smallpox-vaccinia group appear to be identical, 
although the viruses differ in other antigens. 

Extraction of vaccinia virus particles yields at least two other anti- 
gens, a nucleoprotein antigen removable by alkali and the so-called 
X-agglutinogen. Both antigens induce production of antibodies that 
agglutinate the virus particles but do not neutralize them. The an- 
tigens that give rise to neutralizing antibodies apparently cannot be 
extracted from the virus particles without being destroyed. 

Many other viruses give rise to specific soluble substances com- 
parable to the LS antigen of vaccinia in their relation to the viruses. 
For example, influenza viruses are accompanied by specific comple- 
ment-fixing substances (355). These are virus specific, in the sense 
that the SSS of influenza virus A do not react with sera against the 
serologically unrelated influenza virus B. The influenza SSS are not 
strain specific, however; they are apparently identical in all strains of 
a given virus type (A or B). The virus particles themselves contain, 
besides complement-fixing antigens similar to their SSS, other comple- 

CH. 6 Medical Aspects of Virus Serology 131 

merit-fixing antigens, which are often strain specific. Neither of these 
complement-fixing antigens is identical with the antigens that induce 
neutralizing antibodies. 

The soluble antigens play an important role in diagnostic procedures. 
They are often readily obtainable in fairly large amounts, for example, 
from the dermal pulp of vaccinia-infected rabbit skin or from the 
spleen of animals with lymphocytic choriomeningitis. Complement- 
fixation tests may be used to reveal antibodies against soluble antigens 
or to detect the antigen itself in a patient's serum. 

The soluble antigens are also of great theoretical interest. Indeed, 
they appear to represent portions of specific virus material produced 
in excess by the infected hosts. They may originate, either by excess 
formation of specific virus components or by breakdown of virus par- 
ticles within the host. The first alternative seems to be supported by 
the appearance of complement-fixing SSS in eggs infected with in- 
fluenza viruses several hours before any virus activity is recoverable 
(302; 353). This method for investigating virus synthesis by following 
serologically the formation of virus-specific materials will be discussed 
in chapter 13. 

It has been suggested that the SSS may play a role in favoring the 
survival and spread of a virus within the host by acting as blocking 
agents for circulating antivirus antibodies (121). It must be men- 
tioned, however, that a combination of SSS with neutralizing anti- 
bodies has not yet been demonstrated for animal viruses, but only for 
the "ultrafiltrable factors" that accompany certain phages. 

The literature on serology of animal viruses gives the impression 
that their serological specificity is rather variable. Even in the same 
outbreak of a virus disease such as influenza several distinct antigenic 
types of virus may be isolated. This situation may be the result of 
natural selection. Indeed, a high rate of serological mutability would 
provide for a wider range of antibody resistance and would, therefore, 
be advantageous to the virus in its struggle for survival. 


The serological properties of animal viruses provide a basis for much 
of the diagnostic and prophylactic procedures of medical virology. 

Antibodies are produced in an infected organism either in special 
tissues or possibly also in the very tissue in which the virus grows 
(581). The antibodies can prevent the spread of virus from cell to 
cell, but are probably inactive against intracellular virus. Circulating 

132 Serological Properties of Viruses 

antibodies are also ineffectual against the spread of viruses that are 
transported along nerve fibers. By the time the symptoms of a virus 
disease manifest themselves, most of the cells that are going to be 
infected already contain the virus and protect it from circulating 
antibody. This seems especially to be so with neurotropic viruses. 

Because of this, passive immunization, consisting of the injection of 
preformed antibodies, has mainly prophylactic applications, for 
example, during epidemics of measles. With diseases such as measles 
(and probably also poliomyelitis), the blood serum of a large number 
of human adults contains antiviral antibodies. It is therefore prac- 
tical to use for passive immunization the gamma-globulin fraction 
extracted from large pools of human plasma. 

The solid natural immunity that follows recovery from most virus 
diseases is partly due to the persistence of circulating antibodies, which 
may be found 20 years or longer after recovery, for example, in yellow 
fever. There is often a lack of correlation, however, between the 
degree of immunity on the one hand, and the presence and amount of 
neutralizing antibodies in the blood on the other hand (554). Tests 
for local antibody may be of greater significance since the degree of 
immunity is closely correlated with antibody present at the site of 
possible infection, as in bronchial washings in influenza (212). 

The mechanisrii of the long persistence of antibodies against some 
viruses is obscure. Some authors incline to the belief that some virus 
or virus fragment remains permanently in the organism and stimulates 
antibody production (556). Other authors favor the idea that an 
antigen induces a modified pattern of globulin synthesis, which is then 
retained after the antigen has disappeared (112). Some nonviral and 
certainly non-self-reproducing antigens, such as bacterial polysac- 
charides, cause prolonged antibody formation, but the serological 
response is never as persistent as that which follows recovery from virus 

Vaccination against virus diseases is of great prophylactic value. Its 
success well illustrated by the case of Jennerian vaccination against 
smallpox by means of vaccinia virus, or by that of vaccination against 
yellow fever with an egg-passage variant represents one of the major 
conquests of medical science. It was thought at first that the only 
vaccines that gave a satisfactory immunity were those containing active 
virus. This was probably due to the fact that the amount of antigen 
present in inactive virus preparations was usually insufficient to stimu- 
late the production of an adequate antibody level. The purification 
of several viruses in large amounts has permitted the preparation, for 

CH. 6 Medical Aspects of Virus Serology 133 

example, with influenza, of powerful vaccines consisting of virus in- 
activated by treatments, such as formolization or ultraviolet irradiation, 
which preserve the antigenic power. 

In vaccination by means of active virus, a low antigen content in the 
inoculum may be compensated by the production of more virus antigen 
in the vaccinated organism. Various methods are used in different 
cases. Partially inactivated virus vaccines have been employed. They 
may owe their effectiveness to the fact that they contain active and 
inactive virus in proper proportions, so that the level of antibody pro- 
duction under the stimulus of the inactive virus is, presumably, suf- 
ficient to keep the reproduction of the active virus at the nonpatho- 
genic level. Vaccines consisting of a mixture of virus and antibody 
probably act through a mechanism of the same type. Their value will 
depend on the relative chances for migration and diffusion of the virus 
and the antibody from the site of inoculation and on the reversibility 
of the antigen-antibody combination. Some vaccinations may be made 
by inoculating infectious virus by an abnormal route, which prevents 
severe infection but not immunization. In all cases, the choice of an 
adequate vaccine is based on purely empirical criteria of effectiveness. 

The most effective vaccines to date have been those consisting of 
mutant forms of viruses that possess the ability to multiply and cause 
antibody production in a host, for which they are either mildly patho- 
genic or not pathogenic at all. The best examples are immunization 
against smallpox with vaccinia or cowpox virus (from vacca = cow; 
hence the words vaccine and vaccination) and vaccination against 
yellow fever with a virus variant 17D that was isolated after repeated 
passages in tissue cultures and in chick embryos. 

Vaccination is often complicated by the fact that what is considered 
as one virus may occur in a variety of serologically different, often 
unrelated strains. For example, three major serological types of polio- 
myelitis virus have been described. The recognition of such sero- 
logical differences is of paramount importance for any effective sero- 
logical treatment and diagnosis. An effective prophylaxis requires 
vaccines containing at least the most frequent serological types (493). 

Serological procedures provide the most important and often the 
only reliable diagnostic tests for virus diseases, since the actual iso- 
lation of virus from patients is difficult and even in specialized labora- 
tories succeeds only in a fraction of cases. The choice of serological 
method varies from virus to virus (13; 345; see 605 for an extensive 
tabulation ) . 

134 Serological Properties of Viruses 

The virus-neutralization test, in which we determine the reduction 
in infectious titer of a virus preparation after mixing it with the pa- 
tient's serum, is laborious and not too satisfactory, but has till recently 
been the only available test with some viruses such as poliomyelitis 
(542), although complement fixation has been reported (125). For 
viruses that agglutinate red blood cells (see chapter 13) the serum to 
be tested may be examined for its power to inhibit hemagglutination. 
Flocculation tests (agglutination or precipitation) generally require 
too much virus antigen to be of practical use, but might be improved 
"by using as antigens virus-coated bacteria or collodion particles. 
Complement-fixation reactions on the patient's serum are often the 
method of choice, using as antigen either the virus itself or its soluble 
antigens. The antigens need not be purified, but anticomplementary 
Substances, particularly lipids, must generally be removed either by 
extraction or by other methods. 

Skin tests have been devised for a few virus diseases. The Frei 
test for lymphogranuloma consists of intradermal injection of inacti- 
vated virus material. A positive reaction a large papule at the site 
of inoculation is generally present both in infected and in recovered, 
immune individuals. A similar skin test for mumps has been described 

Most diagnostic tests become positive late in the disease or even 
only during convalescence, a fact that reduces their value. Some 
reactions appear earlier than others, however; for example, comple- 
ment fixation with soluble mumps antigen can often be detected in the 
first day of disease. Usually at least two samples of serum are tested, 
one taken very early during disease, the other during convalescence. 
Presence of antibodies against a virus in the first sample may indicate 
a previous history of virus infection and generally speaks against a 
role of that virus in the etiology of the current disease. Appearance 
of antibodies in the convalescent sample is the significant finding, but, 
of course, is more valuable in diagnosis and epidemiological survey 
than from the viewpoint of therapy. When feasible, repeated tests in 
the course of the disease provide the most useful information. 



Environmental Effects 
on Virus Particles 

In this chapter we shall discuss the effects of various environmental 
agencies on the properties of free viruses. The range of detectable 
effects is limited by the number of observable properties of a given 
virus. Some of these are physical properties of the virus particles 
themselves, others are phases of virus-host interaction. Thus, we may 
recognize specific chemical changes of a virus, such as can be pro- 
duced by chemical treatment of tobacco mosaic virus, for example. 
Or, we can recognize a disintegration or an aggregation of virus par- 
ticles by changes in their size and shape. We may observe changes 
in antigenic or in hemagglutinating activity. And finally, we may 
observe losses of infectivity. 

Where the process of infection has been analyzed in terms of in- 
dividual steps of host-virus interaction, as for some bacteriophages, it 
may be possible to recognize virus changes that suppress reproduction 
by blocking one or another of the steps. For example, a virus particle 
may be inactive because of inability either to attach itself to a sensi- 
tive cell, or to penetrate through its wall, or to perform some of the 
later roles required for production of new virus. 

Different environmental agents may inactivate virus particles in a 
variety of ways, by suppressing one or more of their activities. Indeed, 
many of these virus activities have been recognized from the study of 
inactive particles. For example, we know that the attachment of a 
phage to a bacterium is not necessarily followed by invasion and kill- 
ing of the bacterial cells, because we find that by exposure to x-rays a 
phage may be rendered incapable of killing the bacterium, although 
it is still adsorbed by it (665). As a whole, the ability of various agents 
to cause a separation of virus properties is an invaluable tool in virus 
research. One important application of this separation is in the prepa- 
ration of noninfectious virus vaccines. 


136 Environmental Effects on Virus Particles 

Inactivation and disintegration studies on virus suspensions may 
provide some information concerning the structure and composition of 
the viruses. We must admit, however, that the information that has 
been gained from straight inactivation studies, including detailed 
analyses of the kinetics of inactivation, has been altogether disap- 
pointing. This is only an expression of our ignorance of the relation 
between structure and function in the chemistry of complex biological 

The information available on the effect of chemical and physical 
agents on virus particles is different from that available, for example, 
for bacteria, partly because it has been gathered by investigators 
guided by different practical purposes. Problems of spoilage are 
hardly involved in work with viruses. The practical goals of inacti- 
vation studies on viruses are the preservation of virus activity in the 
laboratory, the separation of infectivity from serological activity for the 
productidn of vaccines, and the choice of suitable manipulations in 
the handling of viruses for purification and concentration. A more 
distant goal is the possible application of inactivating agents to the 
suppression of virus infectivity inside the host and to the chemotherapy 
of virus diseases. 

We must be aware of the fact that the stability of virus particles 
exposed to various environments is a "chemical stability," that is, a 
static kind of stability. To alter a virus particle we must change its 
structure. The situation is somewhat different from the one encoun- 
tered, for example, in the study of the stability of bacteria. Here we 
are dealing with an "organismic stability," that is, with a dynamic 
system in a state of continuous flux, into which and from which com- 
ponent elements are continuously incorporated, assimilated, eliminated, 
and broken down. In such a system, irreversible inactivation may 
occur following rather mild environmental changes, since any disturb- 
ance of the dynamic situation may lead to lack of protection against 
new situations produced by the organism's own activities. For 
example, a virus and a bacterium may be equally sensitive to a lower- 
ing of the pH. To inactivate the virus with acid, we must add acid. 
The bacterium, however, may be killed by a change in buffer concentra- 
tion which allows it to accumulate enough acid from its own metabolic 
activity to bring the pH down to the toxic level. Studies on -the stability 
of metabolizing bacteria are more properly to be considered as analo- 
gous to studies on the stability of viruses inside the host cells, a field 
that as yet has barely been developed. 

CH. 7 Environmental Effects on Virus Particles 137 

Very few virus changes have been described that are not accom- 
panied by loss of virus activity. To such nonlethal changes belong 
some changes by chemical treatment of tobacco mosaic virus (624), 
and the reproductive delays produced in bacteriophage by small doses 
of radiation (433). 

Changes in the infectivity of virus particles brought about by exter- 
nal treatments can be classified into three major categories: reversible 
inhibition, inactivation without loss of antigenicity, and disintegration. 
Reversible inhibition of bacteriophage infectivity by combination with 
bacterial debris has been observed. Inhibition of infectivity of many 
plant viruses by a variety of substances, including proteins, enzymes, 
and plant and insect extracts, has been described repeatedly. Some 
has been attributed to a reversible combination between virus and 
inhibitor (43). Usually, however, the inhibition of infection is better 
explained by an effect of the inhibitor on the host cells rather than on 
the virus (281a; 621a). A variety of plant extracts can affect plant 
virus reproduction, even when sprayed on infected leaves as late as 
1 day after inoculation with the virus. 

Inactivation without disintegration of the virus particle is generally 
effected by treatments that do not cause extensive protein denatura- 
tion. Radiation and mild treatments with chemicals such as formalin 
or hydrogen peroxide often produce this result. Protracted or intense 
treatments with an agent that at first only causes a loss of activity lead 
to extensive denaturation and disintegration, often accompanied by 
changes in shape and morphology of the virus particles and sometimes 
by their actual breakdown, as for tobacco mosaic virus or vaccinia 
virus treated with alkali. Occasionally several properties of a virus 
can be suppressed one after the other by progressive treatment with 
one single agent such as ultraviolet radiation (301). The properties 
that are destroyed first are probably those that require the greatest 
degree of structural integrity and are, therefore, suppressed even by a 
slight chemical alteration. The ability to reproduce is generally lost 
very soon, whereas antigenicity is very stable, apparently requiring 
for its suppression a far-reaching alteration of the virus proteins. 

It is important to keep in mind that the effect of any treatment on 
viruses is influenced by such conditions as the composition of the 
medium and the temperature. Sometimes it is difficult to decide which 
agent is actually responsible for the observed virus change, since a 
number of unfavorable conditions all present together may cause more 
damage than would the sum of their individual effects. 


Environmental Effects on Virus Particles 


Temperature effects. In crude preparations or in other media of 
complex organic composition, most viruses are quite stable at room 
temperature; inactivation often becomes measurable at temperatures 
around 50-60 C, the range where denaturation of many proteins pro- 
ceeds at an appreciable rate. Important differences among viruses are 
present, however. Thus, some plant viruses, like tobacco mosaic, are 







Figure 51. The inactivation of viruses by heat, (a) Phage T5 in broth at 
various temperatures. From: Adams (6). (b) Plant viruses in plant sap. I. 
Tobacco ringspot, 50 C. II. Tobacco mosaic, 90 C. III. Tobacco necrosis, 86 
C. IV. Alfalfa mosaic, 62.5 C. From: Price, Arch. ges. Virusforsch. 1:373, 1940. 

practically stable at 65 C and are only inactivated at an appreciable 
rate around 70 C. Aster yellows virus and certain other viruses are 
rapidly destroyed even at comparatively low temperatures. The aster 
yellows virus is inactivated in its insect vector at 32 C (389). Infected 
plants can be cured by immersion in warm water ( 390 ) . The situation 
may be different, however, for such inactivation in the host. For 
example, a bacteriophage quite stable at 50 C is inactivated if phage- 
infected cells are kept at 43 C (432). 

The kinetics of inactivation of viruses at different temperatures is 
exponential. The ratio of the virus activity V to the initial activity V 
diminishes with increasing time t of exposure, according to the equa- 
tion for a first-order reaction: V/V = e~ kt . Deviations from this rule 
are generally due either to aggregation of virus particles or to hetero- 
geneity of their heat sensitivity. The course of inactivation of several 
viruses at different temperatures is illustrated in figure 51. 

CH. 7 Physical Agents (Except Radiation) 139 

Because of the exponential inactivation just described, we cannot 
speak rigorously of a "thermal inactivation point" for a virus, since 
survival always depends on the duration of exposure to heat and on 
the initial virus concentration. Many statements in the literature con- 
cerning thermal inactivation points of viruses reflect the occurrence of 
a narrow temperature range within which the inactivation rate passes 
from low to very high values; that is, virus inactivation generally has a 
high temperature coefficient. The range of thermal stability is that 
temperature range in which the inactivation rate is so low that no 
decrease in activity can be detected. 

The exponential inactivation, expression of a first-order reaction, 
indicates that a constant fraction of the particles undergoes an in- 
activating chemical change in each unit of time and that one such 
change is sufficient to inactivate a particle. The temperature effects, 
according to chemical kinetic theory, reflect the fact that in a popu- 
lation of molecules there is a distribution of molecules with different 
energies. The only molecules that can undergo reaction (with a fixed 
probability) are those that possess at least a certain "energy of activa- 
tion" characteristic for the reaction in question. The frequency of 
molecules whose energy equals or exceeds the activation energy in- 
creases with temperature. The rate constant k as a function of the 
absolute temperature T is given by the Arrhenius equation 

k = Ae~ EIRT [10] 

where R is the gas constant (1.987 calories per mole per degree C); 
A and E are constants characteristic for a given reaction; and E is the 
energy of activation. A monomolecular reaction type results from the 
fact that the rate at which the activated molecules react is slow com- 
pared with the rate at which the equilibrium among molecules of 
different energies is established, so that at each time the fraction of 
molecules with energies greater than the activation energy is a constant 
fraction of the total number of molecules. Thus, the reaction rate 
measures the fraction of molecules whose energy equals or exceeds the 
activation energy necessary for the reaction. 1 

1 The dependence of the reaction rate on the temperature is more fully expressed, 
according to the theory of absolute reaction rates (256a), by the equation 

K - -e - ~r e ~~T e 

n n n 

This formulation expresses the kinetics of two reactions, a first reversible one, in 
which the original molecules become activated, and a second one in which the 


Environmental Effects on Virus Particles 

The validity of equation 10 is tested by the so-called Arrhenius plot; 
log k should be a linear function of 1/T. Such a plot for the denatu- 
ration of tobacco mosaic virus is shown in figure 52. The slope gives 
the value of E/R, from which we can calculate the value of E. Data 






I I 






1/TxlO 3 

Figure 52. The natural logarithm of the rate constants k for denaturation of 
tobacco mosaic virus protein at various temperatures, as a function of the inverse 
of the absolute temperature T (Arrhenius plot). The straight lines indicate that 
the energy of activation is constant over the temperature range included. Upper 
line: 3 mg of virus per ml. Lower line: 6 mg of virus per ml. From: Lauffer, 

Price, and Petre (405). 

for various viruses are given in table 17. The activation energy for 
thermal inactivation of most viruses is high, of the order of 30,000- 
150,000 calories per mole. Such high activation energies correspond to 
high QIQ values. 2 For example, an activation energy of 90,000 calories 
per mole corresponds to a Qi of 50. 

Activation energies of this order are unusual in ordinary chemical 
reactions but are observed for protein denaturation, a fact which sug- 

activated molecules undergo the observed chemical change, In equation 10a, 
K and h are the Boltzmann and Planck constants; AFt, AHt, and AS* are 
the changes in free energy, in total energy, and in entropy in the activation re- 
action (AHt corresponds to E in equation 10); AEt is the change in energy 
at constant pressure; P is the pressure; and AVt is the volume change in the 
activation reaction. Equation 10a explains why the reaction rate may be rapid 
even for reactions with high activation energy, if activation is accompanied by a 
large increase in entropy, so that the total change in free energy is small. Studies 
of reaction rates at various pressures can give information on the changes in molecu- 
lar volume that take place in the transition to the activated state. 

2 The Q 10 ( = fc T+10 A r ) is the ratio between the reaction rates at two tempera- 
tures differing by 10 C. 

CH. 7 

Physical Agents (Except Radiation) 


Table 17. Activation energy for the inactivation of several viruses 

by heat 

E (calories 
per mole) 


Tobacco necrosis virus in plant sap 



Alfalfa mosaic virus in plant sap 



Tobacco ringspot virus (56.5-65 C) in plant sap 



Tobacco ringspot virus (45-56.5 C) in plant sap 



Tobacco mosaic virus (84-95 C) in plant sap 



Tobacco mosaic virus (68-84 C) in plant sap 



Tobacco mosaic virus (denaturation of virus protein) 



Influenza virus A 



Bacteriophage Tl in broth 



Bacteriophage T4 in broth 



Bacteriophage T5 in broth 



Bacteriophage T7 in broth 



Sources: a, Price, Arch. gcs. Vinntfnrsch. 1:373 (1940); b, Lauffer, Price, and 
Petre (405); c, Adams (). 

gests that thermal inactivation results from denaturation of virus pro- 
teins. Indeed, the thermal inactivation of several viruses is closely 
paralleled by the loss of their serological specificity, which is de- 
pendent on the integrity of protein structure. This is true, for example, 
for tobacco mosaic virus, latent potato mosaic virus, and several 
bacteriophages. The role of nucleic acid changes in thermal inactiva- 
tion is difficult to evaluate, since little is known about heat effects on 
nucleic acids. 

Some plant viruses (tomato bushy stunt, tobacco necrosis) are rap- 
idly inactivated by heat at temperatures lower than those which 
denature most proteins (see 43; 405). The energy of activation for 
their loss of infectivity is rather low, and there is a wide range of 
temperatures at which these viruses are inactivated at measurable 
rates. Thermal inactivation at the lower temperatures is not accom- 
panied by loss of serological specificity. One may suspect that in 
these cases thermal inactivation is due either to the destruction of 
some particularly heat-labile portion of the virus or to the thermal 
acceleration of virus inactivation by some other mechanism, possibly 
by combination with some medium component. 

142 Environmental Effects on Virus Particles 

The observations reported above concern virus particles suspended 
in crude media or in relatively concentrated salt solutions. Viruses 
diluted in distilled water or saline solutions are often much less stable 
and become rapidly inactivated even at room temperature. Careful 
studies for various bacteriophages (6) have shown that the concen- 
tration of individual ions is the critical element. For example, coli- 
phage T5 is as stable in presence of 10 ~ 2 M Mg+ + or Ca++ as in 
nutrient broth, whereas in 10 -1 M NaCl the rate of inactivation at 
the same temperature is over one million times faster. Higher con- 
centrations of Na + are protective. The energy of activation is not 
appreciably different in different solutions. Evidently the virus can 
combine with various cations to give reversible combinations, which 
are more stable than the virus particles in uncombined form. The 
inactivation rate observed at a certain ion concentration probably 
measures the fraction of virus present in uncombined form. 

Heat treatment is not a choice method for effecting separation of 
various virus properties, since these properties are often lost at similar 
rates. An interesting exception is shown by the influenza viruses, 
which at 55 C lose their infectivity much faster than their hemaggluti- 
nating ability (327). 

Effects of pH. The pH stability of different viruses is of interest 
mainly in connection with virus extraction and purification. At least 
for tobacco mosaic virus, the stability range of virus activity is nearly 
the same as the range of stability of the virus protein in respect to its 
physical characteristics. At pH above 8.5 the virus is broken into 
smaller units. It has been claimed that these can be rejoined together 
by lowering the pH, although without return of infectivity (587). 

Most animal viruses have optimum stabilities around pH 7, with 
ranges varying widely around this point. Infectivity may be sup- 
pressed by acid or alkaline treatment without loss of antigenicity. 

Drying and shaking. Most viruses withstand drying from the frozen 
state, a procedure commonly used for the preservation of virus speci- 
mens. Wide differences exist in the stability of different viruses in the 
dry state. Inactivation is probably brought about by oxidation in the 
course of the drying process. The presence of foreign proteins in the 
medium (as in blood serum, plant saps, and bacteriological media) 
exerts an appreciable amount of protection, probably due to a com- 
petition for oxidizing agents. 

This is an instance of a very general situation, by which foreign 
proteins protect virus activity against inactivating agents by competi- 
tion mechanisms, just as an excess of one protein protects another 

CH. 7 Chemical Agents 143 

protein against denaturation. Similar protection by proteins is ob- 
served when a suspension of virus (phage) is shaken. Here the for- 
eign protein prevents "surface denaturation" by competing for surface 
area (5). Addition of blood serum or beef extract is a common prac- 
tice in virus preservation. For research purposes, salt-free purified 
gelatin is often used as a protective agent. 


In spite of many studies, it must be acknowledged that little funda- 
mental information has yet come from the study of chemical treatments 
on viruses, mainly because of the difficulty of interpreting changes in 
virus properties in terms of chemical changes. 

Enzymes. Enzyme studies are sometimes complicated by the occur- 
rence of some reversible inhibition of virus activity by combination 
with enzymes just as with many other proteins. No meaningful pat- 
tern has emerged concerning the ability of one virus or another to 
withstand different enzymes. Moreover, viruses may be inactivated 
by nonenzymatic impurities in enzyme preparations. The resistance 
of tobacco mosaic virus and vaccinia to some proteolytic enzymes such 
as trypsin has been utilized in purification procedures to get rid of 
proteinaceous impurities. 

In spite of their relatively large content of nucleic acid, virus par- 
ticles are not inactivated by nucleic acid depolymerases from pancreas 
( ribonuclease and desoxyribonuclease ) . The nucleic acids and the 
proteins of viruses, however, after being extracted from the virus par- 
ticles are rapidly attacked by nucleases and proteolytic enzymes, 
respectively. In their native form within the virus particles, they are 
evidently not accessible to enzyme attack, probably because of in- 
accessibility of the atomic groups upon which the enzymes can act. 
Thus, the organization of virus particles, like the organization of living 
cells but not necessarily by the same mechanism, protects the indi- 
vidual components against enzymatic breakdown. In some bacterio- 
phages, protection against desoxyribonuclease is probably due to the 
protein membrane of the virus particles (308). 

Other chemicals. Several chemicals, when used in the proper con- 
centrations, produce inactivation of viruses without loss of serological 
specificity. Formalin and phenol have been employed successfully in 
the preparation of several noninfectious vaccines. 

Different viruses are inactivated at different rates by the same chemi- 
cal. The inactivation produced by chemical agents may sometimes be 

144 Environmental Effects on Virus Particles 

reversed in vitro by suitable treatments. For example^ formalin in- 
activation of tobacco mosaic virus may be partially reversed by dialysis 
(561) A Some bacteriophages, after inactivation by formalin, can be 
partially reactivated by placing them in a medium with high protein 
content (376). Mercuric chloride inactivation of several viruses has 
been reversed by treatment with hydrogen sulfide. Apparently the 
mercuric ion suppresses virus activity by combining with the sulf- 
hydril groups of proteins, from which it can be displaced competi- 
tively by other SH compounds. Sulfhydril compounds can also 
counteract the inactivating effect of several oxidizing agents, a fact 
which identifies the SH groups of the virus as the sites of the in- 
activating oxidation. In contrast with such reversible inactivation is 
that produced by chemicals, such as urea, that cause extensive protein 
denaturation. Virus particles inactivated by urea are often broken 
down into fragments (405). 

We already mentioned that certain chemical treatments can produce 
chemical changes detectable by analysis of virus proteins without 
suppressing the virus infectivity. Schramm and Miiller ( 588 ) reported 
that all the free amino groups of tobacco mosaic could be covered 
without inactivation, but Stanley and his coworkers (see 624) con- 
cluded that only 70% of these groups (plus 20% of the phenol-plus- 
indole groups ) could be combined with acetyl or phenylureido groups 
without loss of infectivity. The resulting virus was homogeneous in 
size and differed electrophoretically from normal virus. 

Iodine could block the sulfhydril groups of tobacco mosaic with- 
out inactivation, but iodine treatment prolonged enough to affect the 
tyrosine groups was inactivating. The activity of the modified viruses, 
although undiminished when tested in one host, was occasionally re- 
duced in another host, a fact that indicates that virus growth in dif- 
ferent host cells requires different levels of structural integrity. 

From a practical point of view, the observation that most viruses will 
remain active for several years in 50% glycerin, which kills most bac- 
terial cells, has been utilized in preserving virus vaccines. 

Virus inactivation and chemotherapy. Most of the agents that in- 
activate virus particles in vitro are not specific in their action on viruses, 
but act because of their ability to alter proteins. The purpose of 
chemotherapy being the suppression of a virus inside its host without 
damage to the host, any chemotherapeutic agent must have a differ- 
ential toxicity for host and parasite. Protein-denaturing treatments, 
however, generally produce nonspecific protoplasmic destruction. 
They will, therefore, be of little use in therapy. Exceptional cases 

CH. 7 Chemical Agents 145 

may occur. Inactivation of some plant virus diseases by heat treat- 
ment of infected plants is apparently due to the remarkable heat sensi- 
tivity of these viruses as contrasted with the host plants (390). 

Antibacterial chemotherapy has been successful in the use of sub- 
stances, such as antibiotics, chosen on empirical grounds and later 
proved to be inhibitory of enzymatic processes. Chemotherapeutic 
success depends on differential susceptibility of host and parasite, 
either because the host does not require the inhibited enzymes or 
because of different sensitivity or accessibility of host enzymes and 
bacterial enzymes. The classic example is that of the sulfonamides, 
which interfere with the utilization of paraaminobenzoic acid, the re- 
quirement for which is greater for many pathogenic bacteria than for 
animal organisms. 

As we shall see in later chapters, viruses appear to use for their 
reproduction the enzymatic machinery of the host cells. It therefore 
seems likely, on the one hand, that the metabolic-inhibitor approach 
to antivirus chemotherapy will be less likely to succeed than in anti- 
bacterial chemotherapy. On the other hand, success might depend 
on the discovery of enzyme systems, either carried by the host cell or 
possibly by the virus itself, whose suppression would be tolerated by 
the host but not by the virus. Thus, the logical expectation is that a 
successful chemotherapy of virus diseases will be the outcome of a 
detailed analysis of the physiology and biochemistry of the virus- 
infected cell, which may reveal the vulnerable points of the enzymatic 
processes involved. 

There are objections, however, to this reasoning. Our knowledge of 
the synthetic mechanism of cells is still very limited, and a clarification 
of virus-host metabolism may take decades. Indeed, an appreciable 
amount of information on cell metabolism has been obtained from a 
study of the mode of action of empirically chosen antibacterial chemo- 
therapeutic agents, such as sulfonamides and various antibiotics. Clari- 
fication of virus-host metabolism may well follow rather than precede 
chemotherapy's successes. It is a fact that in bacterial diseases most 
of the enzyme inhibitors rationally discovered through metabolic stud- 
ies have proved of no chemotherapeutic value. 

The intensive search for chemotherapeutic agents against virus dis- 
eases has yielded some limited successes. Several infections produced 
by viruses of the psittacosis group in man ( lymphogranuloma vene- 
reum, trachoma) respond to sulfonamide therapy, although psittacosis 
itself does not. Psittacosis, however, is successfully treated by peni- 
cillin, chloromycetin, and aureomycin. Aureomycin, terramycin, 

146 Environmental Effects on Virus Particles 

chloromycetin, and related antibiotics have been used widely in a 
variety of virus diseases, especially in the ill-defined group that goes 
under the name of "primary atypical pneumonia," but the therapeutic 
results are questionable (346). Aureomycin and terramycin have 
proved effective in checking certain virus diseases of mice (257). 

Another possible line of approach to virus chemotherapy stems from 
the discovery of specific organic inhibitors of virus adsorption onto 
susceptible cells. Certain bacteriophages are not adsorbed by sensi- 
tive bacteria in the presence of indole (170). Influenza virus can be 
prevented from infecting sensitive cells by certain polysaccharides 

Formation of tobacco mosaic virus in isolated discs of tobacco leaf 
tissue can be inhibited by purine and pyrimidine analogues (151; 477). 
Amino acid analogues can prevent formation of influenza virus (2), 
but their potential therapeutic value is doubtful. Other approaches 
to virus chemotherapy include attempts to interfere selectively with the 
energy supply of virus-infected cells (199a). 


Radiations occupy a special position among the environmental agents 
that can affect? virus particles. They provide a powerful tool in virus 
research. To understand why, let us recall briefly some fundamentals 
of radiology. 

Electromagnetic radiation includes infrared, visible, ultraviolet, 
roentgen, and gamma rays, each type of radiation being characterized 
by its wavelength. If a beam consists of a single wavelength it is called 
monochromatic. Radiation has a dual character, that of a wave phe- 
nomenon and that of a discontinuous sum of discrete packets of energy 
called quanta. The energy E of a quantum depends on the wave- 
length A according to the relation E = 12,400/A. If A is given in 
angstrom units, E is obtained in electron volts ( 1 electron volt or ev 
is the energy of an electron accelerated by a potential difference of 
1 volt). Yellow light of 5800 A consists of quanta with 2.14 ev; ultra- 
violet light of 2536 A, the major output of common germicidal lamps, 
has quanta of 4.9 ev; for x-rays of 0.1 A, the quantum energy is 
124,000 ev. 

All radiation effects are caused by the absorption of radiation energy 
in matter. They depend on the energy of the quanta and on the nature 
of the absorbing material. Irradiation, then, is a bombardment with 
a beam of quanta. Each electron belonging to an atom may "absorb" 

CH. 7 Radiation Effects 147 

a quantum, and chemical change may ensue if enough energy is 
transferred to the electron. The energy required to pull an electron 
completely out of an atom or molecule (ionization energy) is of the 
order of 10 ev or more. A quantum of x-ray or gamma ray (ionizing 
radiations) can remove an electron from an atom, leaving a positive 
ion (ionization). The ejected electron generally possesses enough 
energy (fast electron) to act as a secondary bullet, ejecting electrons 
from other atoms or molecules until all its energy is spent. The 
ejected electrons will be captured by other atoms to form negative 
ions, so that for every electron that is ejected an ion pair is produced. 
A molecule from which an electron has been ejected has a high proba- 
bility of undergoing chemical change in the course of the resulting 
electronic rearrangements. 

Particulate radiations (beta rays or fast electrons, alpha rays, pro- 
tons, deuterons) act like the fast electrons emitted under x-ray bom- 
bardment. The main differences consist of the more-or-less close 
packing of ionizations along the track of the particles through matter, 
which for heavier particles is practically an "ionization column." 

Weaker quanta (visible and ultraviolet light) do not remove elec- 
trons, but can only produce excitations, that is, they can raise an 
electron to a state of higher energy (nonionizing radiations). The 
rearrangements that follow may lead to chemical change, but the 
probability of chemical change following excitation of complex organic 
molecules is generally much lower. Most excited atoms or molecules 
return to the original condition without chemical change, and the 
extra energy is dissipated into heat. Excitations are also produced by 
ionizing radiations. About one-half of the energy of x-rays is spent 
in producing such excitations. 

There is a fundamental difference between ionizing and nonionizing 
radiations. Absorption of ionizing radiation is nonselective among the 
atoms and molecules exposed, and therefore the probability of ioniza- 
tion of an atom is unaffected by its chemical combination. Thus, the 
absorption ability of a substance depends only on its density and on 
the atomic number of the component atoms. Most biological materials 
differ little from one another (and from water) in average atomic 

Absorption of ultraviolet and visible light depends instead on chem- 
ical structure, since a quantum of a given energy will only be absorbed 
by electrons possessing the proper excitability, which is determined by 
intraatomic and interatomic forces. Thus, among substances of bio- 
logical importance, nucleic acids strongly absorb ultraviolet by means 

148 Environmental Effects on Virus Particles 

of their purine and pyrimidine rings. The absorption coefficient 
( logarithm of the inverse of the fraction of energy absorbed by a 1-cm 
layer of a substance) for various wavelengths defines the absorption 
spectrum of a substance (see figure 48). Absorption spectra of puri- 
fied virus suspensions are easily measured with a spectrophotometer. 

Visible light is absorbed only by colored substances and is generally 
much less effective than ultraviolet in producing chemical change, the 
quantum energy (3 ev or less) being lower than the energy of most 
chemical bonds in organic materials. 

We see then that the characteristic feature of the absorption of 
radiation is the production in the irradiated material of modified atoms 
and molecules distributed in a scattered way. The scattering will be 
geometrically more random in weaker radiations, where the full energy 
of one quantum is used in one single collision, than for ionizing radia- 
tions, where ejected electrons act as secondary bullets producing more 
or less concentrated clusters or columns of ionizations. Because of the 
discontinuous nature of the absorption processes, chemical change will 
be produced in discrete spots, as though the radiation energy brought 
each receiving molecule to a condition corresponding to a large rise in 

The analysis oijthe frequency of a biological reaction ( for example, 
the inactivation of a virus) as a function of the energy, the intensity, 
and the dose of radiation may provide information about the number 
of chemical alterations required, their location, and their nature. More- 
over, the randomness of the bombardment often produces differential 
damage in different parts of a virus particle, resulting in dissociation of 
its various properties. Radiations actually represent a sort of micro- 
surgical tool for the analysis of the dependence of various virus prop- 
erties on the integrity of the virus structure. 

We must distinguish between reactions produced directly, by pri- 
mary absorption of radiation in biological material (photochemical 
reactions), from secondary reactions resulting from the chemical ac- 
tivity of substances produced by the primary reactions in the medium 
that surrounds the virus. 

Ionizing radiation. A virus suspension, irradiated, for example, with 
x-rays, loses its reproductive activity at a rate that depends in part on 
the medium. With dilute or purified viruses in water or salt solutions, 
the inactivation is faster than in crude suspensions and is due mainly 
to toxic agents produced by radiation absorbed in the medium (in- 
direct effects) (239; 408; 444). The toxic agents fall into two cate- 
gories: unstable, short-lived agents, probably OH and H radicals, and 

CH. 7 

Radiation Effects 


more stable ones. These, which probably include peroxides, can be 
demonstrated by adding virus to a previously irradiated medium (666). 
Most of the inaclivation of viruses irradiated in salt solutions is due 
to the short-lived agents (see figure 53). The properties of particles 
inactivated in this way may differ from those treated with the long- 

Tl in broth 
T2 in broth 
o T4 in broth 
a T6 in broth 
A T2 in buffer 


Figure 53. The inactivation of bacteriophages by x-rays, either in nutrient 
broth (direct effect) or in phosphate buffer (direct + indirect effects). Abscissa: 
duration of irradiation. Ordinate: proportion of residual infectious phage. X-ray 
intensity 80,000 roentgens per hour. Data from Watson (665, 666). 

150 Environmental Effects on Virus Particles 

lived agents. For example, bacteriophage inactivated by immersion 
in irradiated buffer is still adsorbed by bacteria, whereas the same 
phage irradiated simultaneously with the buffer is quickly rendered 

These types of inactivation are not directly due to primary photo- 
chemical reactions, but are effects of chemical poisons produced by 
radiation in the medium. They share the properties ( for example, the 
temperature dependence) of chemical reactions among solutes. 

The indirect, secondary nature of this inactivation is shown by the 
fact that, if the virus is irradiated dry, or if an excess of some pro- 
tective substance (gelatin, peptone, tryptophan, thiourea) is added to 
the medium, so that it may compete with the virus for the radiation- 
produced poisons, the rate of inactivation becomes much slower. 

Yet, even in the presence of any amount of protective substance 
there remains a residual, unprotectable inactivation. This is pre- 
sumably due to a direct effect of radiation absorbed within the virus 
itself (444). The rate of direct inactivation (see figure 53) generally 
follows the exponential law 

V/V* = e~ kD [11] 

where V is the ipitial virus titer, V the residual one, and D the dose 
of radiation, usually measured in roentgen units. 3 

The rate of this direct inactivation is independent of the tempera- 
ture and of the time over which the dose is spread. Thus, the process 
has the characteristics of a true photochemical reaction. Equation 11 
is similar to the rate equation for monomolecular reactions. It indi- 
cates that each virus particle is inactivated by one independent event 
or "hit," corresponding to one primary act of absorption (generally 
one ionization). Considerations of this type are the basis of the so- 
called "hit theory" of radiation effects (see 407). 

One virus particle is inactivated by one radiation hit. Let us go a 
little further. If the dose D in equation 11 is measured in number of 
ionizations per unit volume, k becomes the volume ("action volume") 
within which one ionization occurs at the dose for which V/V = e~ l 
(since then kD = 1). This action volume can be measured experi- 

8 One roentgen or r is the dose of x-rays or gamma rays that liberates in 1 cm 3 
of air an amount of ions carrying 1 electrostatic unit of electrical charges. This 
corresponds to about 1.8 X 10 12 ion pairs (one negative and one positive) in 1 cm 3 
of water or of material with the average composition and density of protoplasm. 
Doses of particulate radiations, like protons or alpha particles, are often given in 
"roentgens equivalent physical" or rep. 

CH. 7 

Radiation Effects 


mentally and compared with the volume of the virus particles; several 
results of such comparisons are given in table 18. This analysis is less 

Table 18. The inactivation of viruses by x-rays 



Particle Size 

Dose that 

F/Fo = e~ l 

of Target, 
to Lea (407) 


Phage SI 3 


9.9 X 10 6 


Phage Tl 

50 (+ tail) 

9 X 10 4 


Phage C16 

65 X 95 (+ tail) 

4 X 10 4 



210 X 260 

1 X 10 5 


Rabbit papilloma 


4.4 X 10 4 




2.8 X 10 5 


Tobacco necrosis 


9.4 X 10 6 


Tomato bushy stunt 


6.2 X 10 5 


Tobacco mosaic 

15 X 300 

4.3 X 10 5 


complicated than it may appear. The situation may be compared to 
one in which a man fires many shots at random into a large target 
containing an invisible bull's eye, which when hit rings a bell. If he 
knows the area of the target, the number of shots he has fired at it, 
and the number of times the bell has rung, he can calculate the size 
of the bull's eye. 

Suppose that the dose of radiation that reduces the activity of a 
virus suspension to the level V/V = e~ l (0.37) is 60,000 roent- 
gens, a dose that produces 10 17 ionizations per cm 3 . With this dose 
one ionization will occur on the average in each volume 10 ~ 17 cm 3 . 
This will be the action volume. Similar calculations have also been 
used to estimate the size of genes from the rate of induction of gene 
mutation by x-rays (407). 

Table 18 shows that smaller viruses have generally smaller action 
volumes, that is, are more resistant to radiation. Most authors have 
gone farther and have assumed that the action volume or "target" is a 
real volume, which represents that portion of a virus particle within 

152 Environmental Effects on Virus Particles 

which every ionization is effective, while outside of it the virus is 
wholly unsusceptible. For comparison of the action volumes with the 
actual volume of the virus, several refinements are made (407), cor- 
recting for the waste of ionizations that occur in clusters or in tightly 
packed columns (since one particle can be inactivated only once, a 
cluster of ionizations in one target can only be as effective as one ioniza- 
tion). For small viruses, the calculations give target sizes not too dif- 
ferent from the sizes of the virus particles (table 18), whereas for large 
viruses the targets are much smaller than the particles. According to 
the assumptions of the target theory, the large viruses consist of an 
essential genetic portion, the target, and of a portion where radiation 
damage does not suppress reproduction. For several viruses, the target 
has a volume similar to the volume of the viral nucleic acid (210a). 

Variations of the calculated target size with different radiations have 
been interpreted as due to nonspherical shape of the targets, and have 
been used to estimate the number of elements (supposedly spherical 
and corresponding to genes) in which a target must be resolved to 
account for the discrepancy in effectiveness of different radiations. 
This type of analysis may go farther than is justifiable by our knowl- 
edge both of radiation and of viruses and may, therefore, be mislead- 
ing. We knownow that a virus particle may be inactive (unable to 
reproduce) because of any one of a number of reasons inability to 
attach to a host cell, or to penetrate it, or to reproduce in it. Each 
of these losses of function probably represents a damage to a different 
structure, and the probability that an ionization produces an inacti- 
vating damage may be different from one region of a virus particle to 
another. The main conclusion that we can derive from the analysis of 
the direct effect of ionizing radiations on viruses is that small viruses 
are inactivated by almost every ionization within them, while larger 
virus particles with more complex organization may survive several 
ionizations. Inactivation in all cases results from one successful ioniza- 
tion, not from the summation of the effect of several, as proved by 
the rate of inactivation (equation 11). The parallelism between virus 
size and x-ray sensitivity may be useful in estimating the size of a virus 
by interpolation in cases where other methods are unavailable or 
inapplicable (684). 

Ultraviolet light. The rate of inactivation of viruses by ultraviolet 
light has generally been reported as exponential, obeying equation 11; 
more refined study sometimes reveals deviations from the exponential 
rate (figure 54), deviations whose interpretation is uncertain. Neglect 
ing such deviations, one might conclude that the absorption of one 

CH. 7 

Radiation Effects 


quantum is the hit that inactivates a virus. Not all quanta are effective, 
however; the quantum yield, that is, the ratio effective quanta/absorbed 
quanta is often of the order of 10 ~ 4 to 10 ~ 5 , indicating that most of 

0.02 - 






Figure 54. The inactix ation of bacteriophages by ultraviolet light. Abscissa: 
seconds of exposure to ultraviolet light (2537 A; intensity 1 erg X sec" 1 X mm- 2 ). 
Ordinate: proportion of residual infectious phage. The irradiation and assays 
were carried out under yellow, nonreactivating light. Note that the survival curves 
for the T-even phages are not exponential from the start, but extrapolate to an 
initial titer 1.6 times greater than the one experimentally measured. 

the quanta absorbed by a virus particle either fail to produce perma- 
nent chemical change, or produce changes still compatible with infec- 
tivity (506). The surviving particles may show a delay in reproduc- 
tion, however, as is found with bacteriophages (433). Indirect effects 
through the medium can generally be neglected for viruses irradiated 
with ultraviolet in water or salt solutions, which absorb very little of 
this radiation. 

154 Environmental Effects on Virus Particles 

The action spectrum for inactivation, that is, the quantum yield for 
various wavelengths, has not been determined in detail for any virus, 
but from data on the relative effectiveness of various ultraviolet wave- 
lengths it appears that for several viruses the relative effectiveness 
follows closely the absorption spectrum of nucleic acids. This indi- 
cates that most of the inactivating quanta are absorbed by these 

Quantitative work on ultraviolet irradiation requires exposure of 
virus in nonabsorbent media (water, salt solutions) to prevent screen- 
ing by ultraviolet-absorbing substances such as proteins and other 
organic compounds. Thin layers of opaque suspensions can be used; 
continuous flow techniques have been devised for use in ultraviolet 
irradiation for production of virus vaccines (416). Since glass and 
cellophane are opaque to ultraviolet, open or quartz-covered con- 
tainers must be used. 

Inactivation of some bacteriophages by visible light has been re- 
ported (402). It may be due to a photodynamic action of light through 
the agency of some component of the medium rather than to a direct 
photochemical reaction. Several viruses, indeed, have been shown to 
be inactivated by the photodynamic action of dyes in the presence of 
light (516). 

Radioactive decay. An interesting experiment pertinent to radia- 
tion effects has been made (319) with bacteriophage grown in bac- 
teria containing large amounts of radioactive phosphorus P 32 . Phage 
particles containing many radioactive atoms are unstable and lose 
activity at an exponential rate, as the radioactive decay proceeds. 
On the average, a particle is inactivated by 1 out of 10 P 32 disinte- 
grations. All phage phosphorus is presumably in nucleotide form, 
1 atom per nucleotide. Hence, the above observation suggests that 
there is 1 chance out of 10 that the alteration of a single nucleotide 
in the phage particle renders it inactive. Either some nucleic acid 
portions of the virus are intrinsically nonessential, or there is a limited 
chance for each of several alternate portions to be utilized in repro- 
duction, so that the fraction of effective damages may correspond to 
the inverse number of such alternate portions. 

Separation of virus properties by radiation. Probably every virus 
can be rendered uninfectious by radiation while retaining its anti- 
genicity. Ultraviolet-inactivated virus vaccines have been perfected 
repeatedly for rabies, equine encephalomyelitis and other viruses. 
Moreover, radiation-treated virus may still react with its host cells 
even after losing its ability to reproduce. Such residual interaction 

CH. 7 

Reactivation Phenomena 


has been studied especially with bacteriophages and influenza viruses. 
Ultraviolet leaves to some phages the ability to be adsorbed by host 
bacteria and to kill them, although the phage has become unable to 
reproduce. The nonreproducing but killing phage disrupts the struc- 
ture of the host cell, suppresses the synthesis of host enzymes, and 
interferes with the growth of other phages in the same cell, just like 
active phage. X-rays, acting directly on phage, can render it still 


10 20 30 


3 5 

60 120 


Period of irradiation 

Figure 55. The suppression of various properties of influenza virus by ultra- 
violet light. From: Ilenle and Henle (301). 

adsorbable but nonkilling and noninterfering, probably by suppressing 
its penetration into the cell. 

For influenza viruses (figure 55) increasing doses of ultraviolet 
suppress, in the following order, infectivity, toxicity, ability to interfere 
with other strains, immunizing ability, and hemagglutination capacity 
(301 ). Similar observations have been made with other viruses. 

Reactivation phenomena with irradiated viruses. Radiation-inacti- 
vated phage particles have been reactivated in two ways. First, ultra- 
violet-inactivated particles can be reactivated by causing them to be 
adsorbed to bacteria and then exposing these bacteria to visible light 
(193). This photoreactivation is an example of a general phenomenon 
occurring with ultraviolet irradiated organisms. It occurs to a very 
slight extent with x-ray-inactivated phage. 

The second type of reactivation is observed when some phages 

156 Environmental Effects on Virus Particles 

inactivated by ultraviolet are mixed with host bacteria in such a way 
that each bacterium adsorbs several inactive phage particles. Some of 
the bacteria liberate active phage. The phage has been reactivated 
by interaction among inactive particles within the host cell (multi- 
plicity reactivation; 486). 

Both photoreactivation and multiplicity reactivation will be discussed 
in chapter 9. Similar reactivation phenomena have been described 
for influenza viruses (303). Their possible occurrence must be kept in 
mind, for example, in the use of radiation-inactivated vaccines. 

Intracellular irradiation of viruses. Irradiation of virus-infected 
cells has been performed with a number of viruses. In some cases, 
as with papilloma virus of rabbits, the manifestations of disease can 
be suppressed by doses of radiation much smaller than those needed 
to inactivate free virus. The effect is probably due to a nonspecific 
action on the host cells (240). 

With bacteriophages, no cure of the infected cells by radiation treat- 
ment has been observed, but the ability of the infected cells to liberate 
virus can be suppressed by radiation in a way that suggests a direct 
action on the intracellular virus. The changes of radiation sensitivity 
during the infection are indicative of the evolution of the virus inside 
the cell (64; see chapter 9). Radiations have also been found to induce 
mass lysis and liberation of phage from several strains of lysogenic 
bacteria (454). 



Virus-Host Interaction 
The Bacteriophage-Bacterium System 


Up to this point we have discussed the properties of the material 
particles that represent the static, extracellular form of the viruses. In 
this form, viruses may be considered as in a resting state; no repro- 
duction, no metabolism take place. We must now study how viruses 
perform, that is, how they interact with the hosts in which they 
reproduce. We do not know a priori to what extent the properties 
of intracellular virus resemble those of the extracellular virus particles. 
Inside the host cells, as we shall see, the virus behaves as a cell com- 
ponent, and little is known about the structure (and even less about 
the mode of performance) of any cell component as it exists in the 
cell in the native stateA We know that the cell is not merely a bag 
containing a mixture of Components, but an organized and integrated 
whole. Whatever the mechanisms of organization and integration, 
they may confer upon individual constituents properties that are lost 
or unobservable after extraction and isolation of each component. 
Actually, viruses provide very favorable material for the study of 
cellular integration itself. 

/The central problem in the study of the intracellular behavior of 
viruses is that of their reproductive mechanisms. On the one hand, 
clarification of these mechanisms would give us a better understanding 
of the nature of viruses and of their natural relationships, and, in the 
practical domain, would help us control their propagation. On the 
other hand, information on the mechanisms of viral reproduction may 
throw light on the central problem of biology, that of the reproduction 
of individual specific biological elements, such as genes/) Reproduction 
of cells must be a consequence, a final result, of more elementary 
mechanisms, in which individual carriers of specificity, probably indi- 


158 Bacteriophage-Bacterium Interaction 

vidual molecules, are replicated in identical form, except for mutations. 
Viruses carry specific genetic material. Since they can be studied in 
isolation and then transmitted to new host cells it may be possible to 
trace, qualitatively and quantitatively, the changes in the intracellular 
evolution of the specific viral material. Viruses may indeed be the 
material of choice for the investigation of biological reproduction/^ 

In the interaction between a virus and its host, we distinguish a 
series of phases, including the entry of the virus into the host organism, 
its spread and localization, its reproduction inside susceptible cells, its 
liberation from these cells and secondary spreading to other parts of 
the host organism, and finally, its release from the host and transmis- 
sion to new hosts. From the standpoint of virus biology, each of these 
phases raises problems related to specific virus properties. We wish 
to know how a virus particle acts on the surface of a sensitive cell; 
what part of the virus penetrates, and how; what effects it has on the 
properties of the cell; how the virus reproduces, and by what process 
it is released. Moreover, we ask what determines the ability of a virus 
to grow in one type of cell and not in another, that is, what properties 
make a given animal or plant, or a given tissue, capable of supporting 
reproduction of a given virus. Finally, we inquire as to the ecological 
relations between a virus and its hosts, which allow the continued 
propagation and survival of the virus in nature. 

Virus-host interaction at the cellular level has been studied most 
thoroughly in bacterial viruses. These can be considered as simplified 
models because of the unicellular character of their hosts. We shall, 
therefore, start our discussion with an analysis of bacteriophage-bac- 
terium relation. We must realize, however, that results valid for a 
group of viruses cannot a priori be assumed to be valid for other 
viruses. We already know that viruses are grouped together on 
methodological grounds rather than on grounds of natural relationship. 
Even among viruses active on the same host there are fundamental 
differences. It will be our task to describe what is known in well- 
studied cases and then, from the available material, to attempt what- 
ever generalizations may be justified. 


Even within the group of bacteriophages, it is impossible to gen- 
eralize from one virus to another. We recognize, however, two major 
types of host-virus relationships, the lysogenic and the lytic^ In the 
lysogenic relation, phage is carried by bacterial cells from generation 

CH. 8 The Lytic Type of Phage-Host Interaction 159 

to generation without conspicuous lysis and is occasionally liberated 
into the medium. We can recognize this phage if we have available 
another bacterial strain that reacts to it with visible lysis, so that 
plaques or clearing of cultures may be observed. The lytic condition 
is one in which the phage, after infecting a cell, causes it to lyse 
within a short time, generally before it divides again, with liberation 
of new phage. 

Phage strains that can establish a lysogenic relation with bacteria 
are called temperate phages. Strains that do not establish lysogenicity, 
but regularly lyse the cells they infect, are called virulent. 

The lysogenic condition is probably more important in perpetuating 
phages in nature and also in providing information as to the relation 
of phages to cell constituents and as to their origin. Yet, more is known 
today about the lytic condition, which is easier to analyze/Especially, 
a large amount of our information on virulent phages has oBen derived 
from the study of a few selected systems. One such system, illustrated 
in table 19, consists of 7 phages (Tl, T2, T3, T4, T5, T6, T7), all active 
on a common host bacterium, Escherichia coli, strain B. In the past 
decade a large group of workers has deliberately concentrated on the 
study of these viruses in an effort to clarify the mechanisms of viral 
reproduction (19; 169). 


The broad features of the lytic condition can be visualized in the 
dark-field microscope or, better, in electron micrographs taken at 
intervals after mixing phage and bacteria (448). First, the phage 
particles stick to the bacterial surface. A period of intracellular events 
ensues, followed after a relatively short interval by a sudden lysis and 
liberation of large numbers of phage particles ( see figures 29-31 ) . All 
that is left of the cell structure is a lacerated cell wall, which later 
disintegrates. ) 

The quantitative study of phage-bacterium interaction received its 
greatest impulse by the translation of this cycle of events into a quanti- 
tative procedure called the one-step growth experiment (205). Leav- 
ing the justification of each step for the following sections, we, shall 
describe here the essential procedure (174). 

Concentrated sensitive bacteria under standard cultural conditions 
are mixed with phage, and phage fixation (adsorption) is allowed to 
proceed for a brief period, shorter than the minimum interval between 
adsorption and lysis (the latent period). Then the mixture is diluted 


Bacteriophage-Bacterium Interaction 





J2 8 

S '>T3 




ij i| ilii 

C/3 O 

~ * 1 ^ 1 "a 




a .= 


.s? 2 
1 x 

"i 1 



s .a & 

1 M 2 e 

o ^ 



s| i| | 

j 4> 


x 5 


OH ff^ 


a .a & g 

*c f ^ S rt 2i c 
' *** S< **"* '3 ^ 



a a & -^ 
a 'a S g *, 

^ _ rn 2 cr d 




.2 ^ 

s x 

a .s i? 

a a o bo 
n o c o 

.2 * 

"" "^ "E, 1 

u ^ 







g^ t* w "g ' 
a f c "5 IS 

|ll |l| | 




g * S.f .1 ^ o 





null i 

<! M ^ S O 

CH. 8 The Lytic Type of Phage-Host Interaction 161 

to such an extent that adsorption is practically stopped by the de- 
creased frequency of phage-bacterium collisions. Platings are made 
at intervals for plaque counts, and results of the type shown in figure 
56 are obtained. The low plateau represents the latent period; no lysis 
occurs and the count remains constant because each infected bacterium 








20 30 


Figure 56. A one-step growth experiment with bacteriophage Tl on Escher- 
ichia coli, strain B, in nutrient broth at 37 C. Phage and bacteria, at a ratio 
1:10, were mixed at time 0. The mixture was diluted after 4 minutes, when 45% 
of the phage had been adsorbed. Assays were made at intervals after dilution. 
Average phage yield per infected cell ^ 62 X 100/45 ^ 138. 

that is plated produces only one plaque on the solid medium whatever 
its intracellular phage content at the moment of plating. The rise in 
plaque count corresponds to lysis ( this can be checked by microscopic 
observations), and the final plateau is due to the fact that the newly 
liberated phages fail to meet any bacterium in the diluted mixture and, 
therefore, remain free and do not reproduce any further. This proce- 
dure thus isolates one step or cycle of growth; if dilutions were not 
made, there would be a series of steps blurring into one another, due 
to repeated, nonsynchronized cycles of adsorption and liberation, until 
lysis is completed or until bacterial growth stops, and with it phage 
reproduction. In a mass lysate of this kind, the titer often reaches 
10 11 infectious units per ml, or more. 

The one-step growth experiment provides information on the dura- 
tion of the intracellular phage development and on the amount of 

162 Bacteriophage-Bacterium Interaction 

phage produced. Innumerable modifications can be devised in order 
to study many variables (single, multiple, mixed infection; tempera- 
ture effects; medium composition). Several of them will be discussed 
in the following sections. 

Single particle experiment 



^ 0, 





L. j 

L A 




30 min 


Figure 57. Diagram of a single-burst experiment with phage. In the diagram, 
one out of six tubes (second row) contains an infected bacterium, one tube con- 
tains an unadsorbed phage, and one tube contains an uninfected bacterium. 
After 30 minutes the infected bacteria have lysed. The contents of each tube are 
plated on individual plates for phage count. The first plate (left, bottom row) 
shows the yield of phage from the infected bacterium. From: Delbriick, Harvey 
Lecture Series 41:161, 1945-1946. 

One important modification of the one-step growth experiment is the 
single-burst experiment (108; 205; figure 57). Here, after phage is 
adsorbed by the bacteria, the mixture is diluted until its content of 
infected bacteria is very low, say, 2 per ml; and many small aliquots, 
for example, 0.2 ml each, are distributed into a series of tubes before 
lysis begins. Most tubes will receive no infected bacterium (in the 
above example, a fraction e~ OA = 0.67), some will contain one (a 
fraction 0.4e~ 04 = 0.268), and very few will contain more than one 

CH. 8 The Lytic Type of Phage-Host Interaction 163 

(a fraction 1 - (1 + 0.4)e~ 04 = 0.062). These frequencies are cal- 
culated from equation 1 (see page 49), assuming a random, inde- 
pendent distribution of infected bacteria in the various samples. The 
tubes are then incubated, and after lysis is completed the entire con- 
tent of each tube is plated on one plate for plaque count. Most of the 
plates will have either no plaque or possibly 1 plaque or 2, represent- 
ing residual nonadsorbed phage, which, of course, fails to reproduce. 
The other plates will have relatively large amounts of plaques, and 
most of them contain the yield ("burst") from a single bacterium. The 
burst size distribution is generally wide, showing great disparities in 
the amount of phage produced in individual bacteria (table 20). 

Table 20. The distribution of yields from individual Escherichia coli 
bacteria infected with phage T2 (single-burst distribution) 

(a) Average number of infected bacteria 
per plate, calculated from phage in- 
put 0.73 

(6) Total number of plates 96 

(c) Expected number of plates without 

bursts 96 X e~ 73 = 46 

(d) Number of plates found without 
bursts 39 

(e) Average number of infected bacteria 

per plate, calculated from (d) In (39/96) = 0.9 

(/) Calculated number of plates with 

1 burst 96 X 0.9 X e~ 9 = 35 
(g) Calculated number of plates with 

2 bursts 96(0.9 2 /2)e-- 9 = 16 
(h) Calculated number of plates with 

3 or more bursts 96[1 - (1 + 0.9 + 0.9 2 /2)]e-' 9 = 6 

Plate count distribution 


























































Total count 7938 

Average yield, calculated from (e) 7938/(96 X 0.9) = 92 

164 Bacteriophage-Bacterium Interaction 

In experiments of the one-step or single-burst type, the unadsorbed 
phage can be eliminated before lysis begins, either by washing the 
bacteria free of supernatant fluid or, more conveniently, by adding to 
the mixture some antiphage serum, which inactivates the free phage 
without affecting the adsorbed phage (167). Serum action must be 
stopped by further dilution before lysis begins, to avoid inactivation 
of the newly liberated phage. 


Kinetics of phage adsorption. The fixation of bacteriophage par- 
ticles on susceptible bacteria is easily demonstrated by electron mi- 
croscopy. Quantitatively, it is studied by following the disappearance 
of free bacteriophage from a mixture with bacteria; the bacterial cells 
are separated from the free phage by centrifugation, and the free phage 
is titrated. To avoid liberation of new phage, all measurements must 
be made in the latent period before lysis, which can be prolonged by 
lowering the temperature. Sometimes we can use heat-killed bacteria, 
which retain the ability to adsorb certain phages ( 578 ) ; then the free 
phage can be measured without centrifugation, since phage adsorbed 
irreversibly on heat-killed bacteria is lost and forms no plaque. 

Phage adsorption follows a first-order reaction in relation to the con- 
centrations of both bacteria and phage: 

dPf/dt = -KPfB P f /P = e~ KBt 

where P/ is the concentration of free phage, P is the initial phage con- 
centration, and B is the bacterial concentration. K (cm 3 min" 1 ) is 
the adsorption rate constant. This relation justifies the practice of 
interrupting adsorption by dilution, which decreases the bacterial con- 
centration. From a knowledge of K, we can predict by means of 
equation 12 the amount of phage that will be adsorbed by a given 
number of bacteria in a given time. Equation 12 is valid for mixtures 
with either bacteria or phage in excess, up to ratios "phage"/"bacteria" 
of the order of 200 for large-particle phages. Thus, several phage 
particles can be adsorbed by each bacterium, until complete coating is 

The ratio Pads/B (the average number of phage particles adsorbed 
per bacterium) is generally called the multiplicity of infection. 1 Since 

1 The term multiplicity of infection is something of a misnomer, since the ratio 
P adg /B may be lower than 1. Indeed, expressions such as "multiplicity of infec- 
tion 0.05" are commonly employed. 

CH. 8 Adsorption of Bacteriophage by Bacteria 165 

the adsorption ability of individual bacterial cells is approximately uni- 
form, the fraction B n /B of bacteria that adsorb n particles will follow 
a Poisson distribution: 

B n /B = 


The number B of bacteria without phage will be: 

In typical cases, adsorption of one or more phage particles by a 
sensitive cell results in the death of a bacterium; a colony count of the 
surviving bacteria ( = B ) will, therefore, permit an estimate of P a( is/B. 
This method of measuring the adsorbed phage is particularly useful 
in the study of phage particles which have become unable to reproduce 
but which can still be adsorbed by bacteria and kill them. Phage in- 
activated by ultraviolet light often behaves in this way ( 441 ) . 

Equation 12 is often followed only for the first 90-95% of reduction 
in free phage. A less adsorbable phage fraction remains, which, how- 
ever, after reproducing gives rise to a progeny of normal adsorbability. 
Combination of some phage with inhibitors may be the factor re- 
sponsible for the less adsorbable fraction. 

Chemical requirements for adsorption. The adsorption of phage by 
sensitive bacteria is influenced by the physiological conditions of the 
bacteria (mainly through the smaller size of older cells, and possibly 
through the production of slime or capsules ) and, even more markedly, 
by the composition of the medium. In pure distilled water or in a 
concentration 10 ~ 4 M of monovalent ions (Na+ or K+), adsorption 
of phages of the T group fails to occur. Increasing cation concen- 
trations render adsorption possible; the optimum concentrations, that 
is, those permitting maximum adsorption rates, are characteristic for 
each phage and each ion. Concentrations above optimum often reduce 
the adsorption rate of phage. This is probably due to a combination 
between the ions and the bacterial surface receptors. Some data are 
given in table 21. 

Adsorption in salt solutions with low concentrations of cations is 
reversible, either by dilution or by transfer to distilled water. At higher 
ionic concentrations reversible adsorption is promptly followed by a 
fast irreversible reaction. The reversible adsorption does not lead to 
any detectable changes in the host cell (250; 535). 

In complex salt-containing media or in buffers with optimum ion 
concentration, the adsorption rate constant K approaches values of 4 

166 Bacteriophage-Bacterium Interaction 

Table 21. Optimum salt concentrations for irreversible adsorption of 
the T phages onto Escherichia coli, strain B 

The concentrations given below allow phage adsorption at the maximum rate. 
Data supplied by Dr. T. T. Puck. 

Tl 10~ 3 M Mg+ + or Ca++; or 10~ 2 M Na+ or K+ or NH 4 + 

T2, T4, TO 10- 1 M Na+ or K+ or NH 4 + 

T3 10-* M Na+ or K+ or NH 4 + 

T7 10- 2 M Na+ or K+ or NH 4 +; or 10~ 3 M Mg++ 

to 5 X 10 ~ 9 cm 3 min" 1 . This is approximately the value that can be 
calculated for the rate constant, on the assumption that almost every 
collision between phage and bacterium results in adsorption, from the 
known values of the radius r of the bacteria and of the diffusion 
constant D of the phage (K = 4?rrD; see 165). 

This result, showing the enormous efficiency of the adsorption re- 
action, suggests that phage adsorption can occur following contact of 
the phage with any portion of the bacterial surface. Indeed, the maxi- 
mum absorption capacity of 300 particles of a phage about 100 m/x in 
diameter corresponds to the coating of about one-third of a bacterial 
cell 3 X 1 /A. 

The remarkable fact is that, although specific for each phage, the 
ion requirements for reversible adsorption are not specific for phage- 
bacterium combination. They are identical to the requirements for 
adsorption of phage particles to cation exchange resins or to glass ( for 
example, to sintered glass filters), an adsorption which, biologically 
speaking, is clearly nonspecific (535). 

A similar situation holds for another category of requirements for 
phage adsorption, namely activation by cof actors ( see 17 ) . Phages T4 
and T6, among the strains of the T group, are only adsorbed by bac- 
teria (or by glass filters) if there are present in the medium certain 
cofactors. L-tryptophan is the most effective cofactor of adsorption, 
other amino acids are much less effective. Substitutions in the aro- 
matic ring of tryptophan (for example, in 5-methyltryptophan ) may 
preserve cofactor activity; substitutions in the aliphatic chain abolish 
it. Indole, a tryptophan analogue, is a competitive inhibitor of trypto- 
phan action (170). 

The kinetics of tryptophan action is complex. The amino acid com- 
bines reversibly with phage, the reaction being rapid in both direc- 
tions. The influence of concentration on activation and deactivation 
is best interpreted by assuming that each tryptophan molecule is ad- 

CH. 8 Adsorption of Bacteriophage by Bacteria 167 

sorbed independently onto the phage particle, and that activation 
results from the formation on the phage surface of self-stabilizing 
complexes, each complex consisting of 5 molecules of the amino acid 

Together with the results on the influence of ions, the cofactor 
phenomena can be interpreted by assuming that the phage's surface- 
modified by the cofactor molecules, when these are needed provides 
a base on which ions become attached to give the proper electric 
charges, which must be present to permit contact between the phage 
and the specific bacterial receptors. Irreversible adsorption or fixation 
will then occur by other more specific, possibly enzymatic binding 
mechanisms (250). 

The activity of indole as a tryptophan inhibitor is remarkable, in 
that it provides an example of a potential specific defense mechanism 
against viruses (170). Indole is a product of the metabolism of trypto- 
phan by Escherichia coli, the host bacterium for the tryptophan-requir- 
ing phages T4 and T6. If tryptophan is present in excess in the 
medium indole is produced and can act to prevent adsorption of these 

The nature of the bacterial receptors. As a rule, only bacteria that 
can grow a phage will adsorb it. Some cases are known in which the 
phage can be adsorbed by bacteria unable to support its growth and 
to be lysed; serological cross-reactions have been found between the 
normal host for a phage and the nonhost bacteria that adsorb that 
phage (545). This suggests a role of the antigenic constitution of the 
bacterial surface in phage adsorption. A fairly good correlation exists 
between the possession of certain antigens and the sensitivity to dif- 
ferent bacteriophages ( see 101 ) . For example, the phages active on 
bacteria of the genus Salmonella can be divided into several groups, 
some active on rough strains only, some on smooth ones; smoothness 
and roughness of bacteria depend on the possession of certain antigens. 
Some phages against typhoid bacilli are specific for strains carrying 
the so-called Vi antigen, which is present in virulent strains. These 
phages and their substrains (see page 206) are used in the phage- 
typing method for the identification of individual Vt-strains of 
Salmonella typhosa (157; 157 a}. Phage typing is used with a variety 
of other organisms (156). 

It seems plausible that the surface antigens of the bacteria are 
actually involved in the adsorption of phage. Sensitive bacteria yield 
fractions that inactivate homologous phage specifically, and these 
phage-inactivating agents, once purified, are similar to known bac- 

168 Bacteriophage-Bacterium Interaction 

terial antigens (206; 263; 415). The situation is by no means simple, 
however. Some workers report specific phage-inactivating power in 
polysaccharide fractions (206), others only in carbohydrate-lipoid-pro- 
tein complexes in native form (484) or in their lipocarbohydrate por- 
tion (2,57 a). Also, phage combines specifically with carbohydrate-free 
cell walls obtained by extracting the cells in a series of successive treat- 
ments (668). These residual cell walls are slowly destroyed upon 
incubation with phage, but the mechanism of this process is as yet 

These facts and the numerous exceptions to the rule of similarity 
between antigenic properties and susceptibility to phages may be 
accounted for by the suggestion (109) that phage receptors and anti- 
body receptors are not necessarily identical, but may represent vari- 
ously overlapping portions of certain reactive patches of the bacterial 


Problems of phage reproduction. Following phage adsorption there 
is an interval before lysis during which, normally, more phage is 
produced. The average yield of phage per bacterium or burst size, 
as determined by one-step growth experiments, is the ratio 

(Final titer) (Unadsorbed phage) 
(Initial titer) (Unadsorbed phage) 

The average yields for the T phages on young growing bacteria are 
shown in table 19; yields from old bacteria are lower. 

The series of events between phage adsorption and lysis has to be 
reconstructed from more or less indirect evidence obtained by a variety 
of methods. The evolution of the virus-infected bacterium in the 
latent period is complex. Some stages can be recognized from the 
existence of abnormal situations in which the evolution of the system 
stops early. Thus, phage adsorption may not be followed by any 
other noticeable changes in the bacterium, which goes on dividing; this 
occurs, for example, with x-ray-inactivated phage (665). Or the 
bacterium may be killed but may fail to produce any active phage 
and to undergo lysis, as with ultraviolet-inactivated phage (441), or, 
for phages like T5 and others, in the absence of Ca + (7). 

A phenomenon, which contributes only indirectly to the understand- 
ing of the normal phage-bacterium evolution, is lysis from without 
(166)'; this is produced by rapid adsorption of a very large number of 

CH. 8 The Latent Period of Intracellular Development 169 

phage particles, active or inactive. This lysis apparently results from 
massive damage to the bacterial surface; no active phage is released, 
not even the original ones, and the cell does not undergo the specific 
cytochemical changes which, as we shall see, are often associated with 
phage-cell integration (445). 

Phage reproduction, when present, takes place inside the bacterium. 
The new phage is not visible by electron microscopy on the bacterial 
surface, nor is it accessible to antiphage serum action. The phage- 
reproducing mechanism becomes refractory to antiphage serum action 
within seconds after phage adsorption. Direct observation of the phage 
inside the bacterium, either by dark-field microscopy or by electron 
microscopy, yields results difficult to interpret and of little value; 
indirect methods must be employed. Before describing and evaluating 
such methods and their results, let us consider briefly what we are 
looking for and what we may expect to find. 

A virus particle infects a cell, and several minutes later several hun- 
dred particles emerge. What is the relation of the initial particle to 
the final ones? Does the initial particle penetrate the bacterium as a 
whole or in part? Does it appear as such among the progeny particles? 
Or does it contribute material to the composition of one or more of 
them? If so, are these materials specific? 

How does the initial particle preside over the formation and speci- 
ficity of the new ones? Does it cause a transformation into virus of a 
ready-made precursor, or a synthesis of virus from nonspecific building 
blocks? Are these supplied by the bacterial cell, by the external 
medium, or by both? Are they synthesized and put together by 
mechanisms provided by the bacterium, by the virus, or by both? 
What types of mechanisms are operative: enzymatic, autocatalytic, 
"patternlike," or others? And where does the necessary energy come 

Does a virus particle "grow," that is, produce more viral material 
within its own borders, then split into two or more particles? Or 
is the new virus put together from scratch and all at once, by copying 
a preexistent virus that provides a model for assembly of nonspecific 
blocks? Or else are there intermediate levels of organization in which 
the building blocks, still free in the cell, become more and more virus- 
like through a series of stages of increased structural complexity and 

It is clear that these problems deal with the functional organization 
of the virus-infected cell, and that little can be learned about them by 
analogy with other situations, since the functional organization of the 

170 Bacteriophage-Bacterium Interaction 

cell, particularly at the level at which specific syntheses occur, remains 
as yet a field where even speculation has hardly dared venture. Virol- 
ogy may ultimately prove the field of choice for its clarification. 

The sections that follow provide evidence of various types on intra- 
cellular phage development. They suggest a complex and by no means 
final picture, in which the phage particle introduces a specific portion 
of itself into the host cell, modifies the pattern of specific syntheses in 
the host, and utilizes, at least in part, the enzymatic machinery of the 
host to obtain building blocks, on which it impresses its own viral 
specificity. Most of this development takes place in stages. Mature, 
infectious phage particles are formed only as an end product of the 
process as a whole. 

Premature lysis. Artificial breakage of phage-infected bacteria long 
failed to yield any active phage. Success was finally obtained by the 
use of a variety of methods: sonic vibration (for phages such as T3, 
that can stand this treatment) or "lysis from without" of the cells by 
an excess of a different phage (182); mechanical disruption of cells by 
decompression (234). Cyanide or some other metabolic inhibitor is 
generally added in order to stop the progress of phage development at 
a precise moment. 

These methods reveal that increasing amounts of active phage are 
obtained from infected bacteria in the second half of the latent period 
(figure 58); in the first half, however, no active phage is recovered. 
Thus there is an eclipse of the infecting phage, a phenomenon that 
probably occurs for all virus infections, as shown by the nonrecover- 
ability of most viruses immediately after infection. The eclipse is cer- 
tainly real, and, as we shall see, the first virus particles that appear 
after the eclipse are not the infecting particles; they are already part 
of the final crop of new virus. 

The rate of appearance of infectious, intracellular virus is nearly 
linear. An apparent increase in rate at the beginning is caused by an 
inhomogeneity in the time of appearance of the first active phage 
particle in individual bacteria. The rate of appearance of the new 
virus, however, does not represent the actual rate of phage synthesis, 
but only the rate at which individual phage particles "mature," that is, 
graduate to the infectious stage. 

This is shown, for example, by the effect of addition of the dye pro- 
flavine (diamino acridinium sulfate) to bacteria infected with phage 
T2 (227). The dye allows lysis to occur normally, but only as many 
mature, active phage particles are liberated as were already present 
when the dye was added; none appears if the dye was added before 

CH. 8 The Latent Period of Intracellular Development 171 

the middle of the latent period. If one examines the materials liberated 
by these bacteria, one finds by electron microscopy that they contain 
many inactive, tailless, incomplete phage particles or "doughnuts" 
(figure 28, page 68), in numbers approximately similar to those of 


At 10.5 minutes, sample 
contained < 0.01% 


5 10 15 20 25 30 35 40 45 

Minutes of incubation in growth medium at 37 

Figure 58. Premature lysis of Escherichia coli cells infected with phage T4r. 
Premature lysis was produced at intervals by diluting in a mixture of 10 ~ 3 M 
cyanide ( to stop further phage production ) and of phage T6 in excess ( to induce 
lysis). Abscissa: time interval between infection and dilution in cyanide. Ordi- 
nate: amount of phage present in per cent of the phage yield that would be ob- 
tained by natural lysis. Curve 1 : data from a single experiment, with 7 T4r par- 
ticles adsorbed per cell. Curve 2: combined data from four experiments with 
single infection. Curve 3: control one-step growth curve; single infection. Note 
that the samples taken before 13 minutes contained less infectious phage than the 
original inoculum, From: Doermann (182). 

active phage particles that would have been produced in the absence 
of the dye (175). These "doughnuts" are not adsorbed by bacteria. 
They contain little or no nucleic acid; they contain an appreciable 
amount of sulfur, presumably in protein. They fix complement 
with antiphage antiserum but do not combine with phage-neutralizing 

The doughnuts produced in proflavine are similar in morphology to 
immature phage forms that can be seen in mechanically disrupted 

172 Bacteriophage-Bacterium Interaction 

bacteria before the mature phage particles appear (417). Counts of 
the doughnuts and of tailed particles from bacteria disrupted at various 
times reveal an increase in doughnuts, followed by an increase in tailed 
forms. This suggests a transformation of the doughnuts into tailed 
particles. In the presence of proflavine this transformation fails to take 






g 40 


i i i 


40 60 80' 

Minutes after infection 

160 180 

Figure 59. Increase in the number of immature phage particles ( "doughnuts" ) 
inside bacteria of E. coli B infected with phage T2 in presence* of 4 n of pro- 
flavine per ml. Electron microscope counts on extracts of prematurely lysed 
bacteria. From: DeMars et al. (175). 

place (figure 59). Thus the doughnuts appear to be organized pre- 
cursor elements of the phage particles. At a later stage of evolution 
one detects nucleic acid-free particles, which are still noninfectious but 
are adsorbed by bacteria (464). 

It is clear that phage synthesis goes through stages, the active phage 
particle being a sort of final, mature form. Once in this form, as we 
shall see later, a phage particle no longer multiplies in the bacterium 
in which it has been produced. Ability to infect a new bacterium is 
apparently acquired by a transformation of the particle, which stops its 
reproduction within the cell of its origin. The rate of appearance of 

CH. 8 Transformation of the Infecting Phage 173 

active phage, as obtained from premature lysis experiments, probably 
represents the rate at which the final maturation reaction takes place, 
and may differ from the rate at which the production of individual 
phage replicas takes place. 

The incomplete, tailless particles are not the only phage-specific 
material obtainable, besides active phage, inside infected bacteria. A 
serologically specific ultrafiltrable factor, which combines with phage- 
neutralizing antibody (see page 126), can be obtained by disruption 
of infected bacteria throughout the eclipse period (175). This factor 
increases in amount prior to and in parallel with the increase of active 
phage. Some of it is liberated, along with mature phage, upon normal 
lysis. The ultrafiltrable factor probably represents material on its way 
to becoming phage and containing the antigen responsible for neu- 
tralizing antibody. 

Thus premature lysis experiments tell us that, following infection, 
the infecting particle is so altered as to be noninfectious. After an 
eclipse period lasting almost half the latent period, new phage-specific 
elements of various complexity can be detected in increasing number, 
until the mature particles are finally formed. 

Transformation of the infecting phage following adsorption. For 
phage T2 infecting E. coli B, the disappearance of the infecting phages 
as infectious units and the meaning of the eclipse period have been 
explained by the brilliant work of Hershey and Chase (313), using 
isotopic tracers. Phage labeled with P :< - in its nucleic acid (or with 
S :i5 in its sulfur-amino acids) can be obtained by growing it in labeled 
media and purifying it from the lysates. This phage can then be used 
to infect unlabelcd bacteria, and the fate of the P or S of the infecting 
phage can be traced by measuring the radioactivity of various fractions 
(whole bacteria, extracts, medium, etc.). 

Using labeled phage T2, Hershey and Chase found that the phos- 
phorus of the infecting phage, that is, its nucleic acid, separates from 
the sulfur. The phosphorus apparently penetrates the bacterial cell 
and acquires some of the properties of the bacterial nucleic acid; for 
example, it becomes accessible to desoxyribonuclease action in bac- 
teria heated at 80 C. Surprisingly enough, the sulfur appears instead 
to remain at the surface of the bacteria. It can be removed from the 
bacteria by violent shaking of infected bacteria in a rotary mixer 
(Waring Blendor) without preventing the growth of phage in these 
bacteria and its normal liberation. 

This remarkable result can be visualized in the light of the known 
organization of the particle of this phage, which, upon osmotic shock 


Bacteriophage-Bacterium Interaction 

(transfer from 5 N to 1 N NaCl), separates into a S-rich, P-free "ghost" 
or "skin" (see figure 28, page 68) carrying the main antigens of the 
phage, and a fraction of highly polymerized nucleic acid (308). Ap- 
parently, upon infection, the phage skin remains on the bacterial sur- 
face, whereas the nucleic acid, possibly accompanied by some S-free 
protein, penetrates and initiates the production of the new phage 
( figure 60 ) . These observations on phage T2 have been confirmed for 
several other coli-phages of the T1-T7 system ( see, for example, 401 ) . 










Figure 60. Diagram of the transformation of the infecting phage particles ac- 
cording to data by Hershey and Chase (313). 1. Phage and bacterium. 2. Ad- 
sorption. 3. Separation of the phosphorus-containing core ( P, stippled ) from the 
sulfur-containing skin (S). 4. Removal of skin by stirring in a Waring Blendon 
5. Production of T&ew phage, containing phosphorus and sulfur without participa- 
tion of the skin of the infecting phage. 

These observations explain at once the eclipse of recoverable infec- 
tive particles following infection and the inability of antiphage serum 
to stop phage reproduction. Complete phage particles cease to exist 
as organized units, and the phage skins, carrying most and possibly 
all the phage antigens, remain on the outside and take no part in the 
reproductive process. 

The eclipse is thereby explained. The problem of phage production 
becomes that of the mechanism whereby the sulfur-free phage pri- 
mordium or "nucleus" that penetrates the cell gives rise to the full, 
mature virus particles in the yield. 

The discovery of hydroxymethyl-cytosine in phages of the T2 group 
(687b) has made it possible to follow the synthesis of phage DNA in 
infected bacteria (315a). When the phage-specific DNA and host- 
specific DNA are measured at intervals in infected cultures, the phage 
DNA begins to increase soon after infection, while the bacterial DNA 
decreases. After the appearance of mature phage the phage DNA 
continues to increase and is always in excess over the amount ac- 

CH. 8 Origin of Phage Constituents 175 

counted for by the infectious particles. The excess phage DNA corre- 
sponds to the amount needed to supply 40-80 phage particles. If this 
phage DNA is in the form of immature phage "nuclei," we may con- 
clude that during growth of phage T2 in a bacterium there is a popu- 
lation consisting of up to 40-80- immature particles. This population 
is presumably repleted by multiplication and depleted by maturation, 
and its size remains more or less constant. The multiplying noninfec- 
tious form of phage has been called the vegetative phage. 

Thus, the phage DNA penetrates the bacterial host, while little or 
no phage protein does. Phage DNA increases very early after infec- 
tion, while phage-specific protein is not observed, even by sensitive 
serological tests, until a few minutes before mature phage appears. It 
is therefore puzzling that the very first morphologically recognizable 
new phage elements that can be isolated from prematurely disrupted 
bacteria are DNA-free, antigenically specific particlesthe doughnuts 
(417) or their tailed successors (464). It is possible that the phage 
proteins are produced around the phage DNA, but that they remain 
separable from it until a late reaction occurs, which binds them into 
stable form and puts the last touch upon the mature particle. 

Origin of phage constituents in infected bacteria. The amount of 
new phage produced in a bacterium sometimes corresponds to as much 
as one-tenth or more of the mass of the bacterium. What is the 
origin of the materials that go to form the new phage? How are they 
synthesized? And where does the energy for their synthesis come 
from? Phages seem to contain only proteins and DNA. These sub- 
stances could either derive from constituents of the bacterial proto- 
plasm already present in the cell at the time of infection, or they could 
be assimilated from the medium after infection. Experiments with 
isotope tracers show that both sources contribute materials for phage 
synthesis. These experiments are carried out as follows (143): 

Suppose we grow bacteria in a medium containing phosphorus com- 
pounds "labeled" by a certain content of the radioactive isotope P 32 
(one-half of whose atoms disintegrate in every 14.3-day interval, giving 
rise to S 32 atoms). All the phosphorylated components of the bacteria 
will contain some phosphorus recognizable by its radioactive emission 
of beta rays. These bacteria are washed, resuspended in nonradio- 
active medium, infected with phage, and allowed to lyse. The new 
phage is purified, and the radioactivity, that is, the proportion of P 32 
in its phosphorus, is measured. If 25% of the phage phosphorus is 
derived from material present in the bacteria at the time of infection 

176 Bacteriophage-Bacterium Interaction 

(bacterial contribution) and 15% comes from phosphorus taken in from 
the medium after infection (medium contribution), the ratio (P 32 /P 
total) in the new phage will be one-fourth the ratio in the labeled 
medium in which the bacteria had been grown. Thus the P 32 content 
of the new phage will be a measure of the amount of phosphorus that 
has been transferred from bacterial cell to phage. Of course, the 
"medium contribution" need not represent material that goes directly 
into phage. It is simply the material that has been taken up by the 
cell after infection. 

Alternatively, unlabeled bacteria may be placed in labeled medium 
at any desired interval of time before or after infection, and the trans- 
fer of P 32 from medium to phage may be measured. Or else, labeled 
phage, prepared by growth in labeled bacteria in labeled medium, may 
be used to infect unlabeled bacteria in unlabeled medium, to study 
the fate of the phosphorus of the infecting phage, as already discussed 
in the preceding section. Similar experiments can be done by labeling 
sulfur, carbon, or nitrogen with isotopes, either radioactive or stable; 
the stable isotopes are measured with the mass spectrograph. Since 
carbon and nitrogen are present in both nucleic acid and protein, they 
can be used to label either moiety, whereas phosphorus provides in- 
formation only about the origin of nucleic acid, and sulfur only about 
that of sulfur-containing protein. We can also incorporate into bac- 
teria certain preformed, labeled building blocks (purines, pyrimidines, 
amino acids), whose transfer from host to virus we can then follow. 

The main results obtained from this type of study on phages of the 
T1-T7 group can be summarized as follows (144; 237; 313; 384; 392; 
463; 465^ 627 a; 667 a) : 

1. Of the phosphorus of the infecting phage, about 30-50% appears 
in the new phage; whether within one of them or distributed among 
many we do not know. The new phage in turn transfers the same 
proportion of its phosphorus to a following phage generation. With 
a yield of 100 particles per bacterium, the contribution of one infecting 
phage particle is not more than 1 /{>oo of the DNA of the new phage. 
Most of the parental phosphorus is found in the progeny produced in 
the first half of the latent period. 

2. No sulfur is transferred from parent phage to progeny phage, as 
expected from the fact that the phage skin, containing all the phage 
sulfur, takes no part in reproduction. Likewise, no significant amount 
of amino acids is apparently transferred from parent phage to progeny 

CH. 8 Origin of Phage Constituents 177 

3. About 20-35% of the nitrogen and of the phosphorus in the DNA 
of the new phage in T2, T4, T5, and T6 conies from the bacterial con- 
tributions; the rest (65-80%) comes from the medium. For T3 or T7, 
about 75% of the phosphorus comes from the host, and for Tl about 

4. Purines and pyrimidines present in host DNA contribute 15% or 
more of the corresponding compounds in the new phage. 

5. About 80-90% of the nitrogen of the phage protein is contributed 
from the medium, 10-20% from the host. 

6. The phage protein nitrogen is derived from substrates in rela- 
tively rapid equilibrium with the medium, whereas the nitrogen and 
phosphorus in phage DNA are derived from a fairly long-lived pool 
of compounds (purine or pyrimidine bases, nucleotides, polynucleo- 
tides). This pool is fed partly by preexistent bacterial protoplasm, 
partly by new syntheses from the medium. The contribution of the 
infecting phage to the phosphorus of the new phage may either go into 
the pool or represent a transfer of intact chunks of phage DNA. 
Definite evidence on this point is still lacking. The bacterial con- 
tribution to the phosphorus pool derives predominantly from the bac- 
terial DNA, but the ratios of various DNA constituents are not identi- 
cal in bacterial and in viral DNA. This indicates that the bacterial 
DNA is not transferred in intact large chunks to the new phage. 

7. By making labeled medium available at various times for 1-minute 
intervals only, preceding or following phage infection, and isolating 
the phage obtained after premature lysis at various times, it can be 
shown that each phage particle draws, from the available pool, ma- 
terials that have entered the pool at any time during an interval of 
several minutes preceding its maturation. The average time between 
the uptake of a phosphorus atom from the medium and its appearance 
in mature phage is about 14 minutes; for sulfur atoms, about 7 minutes. 

8. The phage particles that are produced last contain proportionally 
a higher medium contribution of phosphorus and pyrimidines than the 
particles formed first, as though the original bacterial contribution had 
become depleted. 

These results should be viewed together with the following evidence 
obtained by chemical analysis of phage-infected bacteria (144; see fig- 
ure 61). After infection with phage T2 the total DNA in the cells 
remains approximately constant, although phage DNA is already 
formed during this period (315a). Later, DNA increases at a rate 
about linear. The total RNA remains constant. The total protein 
increases linearly from the moment of infection (144). 


Bacteriophage-Bacterium Interaction 

The data as a whole fit the following picture of the synthesis of 
phages like T2 and its relatives. Phage infection stops the synthesis of 
bacterial RNA. It allows or stimulates a breakdown of bacterial DNA; 
it stimulates synthesis of new DNA and of new protein, both marked 
for use in phage synthesis. The phage protein derives mostly from the 
medium and from low-molecular-weight, acid-soluble nitrogenous frac- 
tions present in the cell at the moment of infection. The DNA derives 















Figure 61. Schematic representation of the synthesis of various constituents of 
Escherichia coli, strain B infected with phage T4r. Infection takes place at time 0. 
a, The amounts of various chemical fractions; b, the relation of DNA synthesis to 
phage formation. Data from Cohen (143, 144). 

from a pool fed by the medium and by breakdown of the bacterial 
DNA. 2 

Bacterial enzymes and the synthesis of phage constituents. We have 
seen that the phage-infected cell performs syntheses which lead to the 
formation of phage-specific materials. Several questions arise. Which 
portion of the new syntheses is carried out by bacterial machinery, and 
which by phage machinery? Is there synthesis of specific bacterial 
components or only of specific phage components? How are these 
syntheses directed? 

Extensive evidence indicates that in bacteria infected with the T 
phages many bacterial enzymes that were present at the time of infec- 
tion continue to work; they provide energy and synthesize nonspecific 

2 An older idea, according to which the phage simply catalyzed the conversion 
of ready-made precursor into active phage (385), is incompatible with the results 
of the tracer analysis. 

CH. 8 Role of Phage in Synthesis 179 

building blocks, but these are used to build phage protoplasm rather 
than bacterial protoplasm. We use the word "protoplasm" as a con- 
venient term to include all that is chemically and functionally specific 
in a given cell (enzymes, antigens, genetic substances), even though 
the chemical basis of the specificity still escapes us. We may list here 
some of the evidence for the above statement (see 144) : 

1. After phage infection, bacterial respiration continues at the same 
rate as at the moment of infection. No rate increase occurs, even with 
bacteria that were actively increasing in size and in respiratory activ- 
ity before infection. 

2. Inhibitors that block the energy-yielding process of bacteria also 
prevent phage production. However, pretreatments that block cell 
division but allow bacterial metabolism to continue (penicillin, mus- 
tard gas, radiation) allow phage growth to take place (18; 534). 

3. In bacteria that require for growth a supply of some essential 
building block for example, an amino acid phage synthesis requires 
the same supplement (541). It would be interesting to test whether 
this rule holds only for substances, such as amino acids, that actually 
appear in the phage protoplasm. Continued requirement for building 
blocks not present in phage particles would suggest that synthesis of 
some constituents of cell protoplasm takes place in phage-infected cells. 

4. No enzymatic adaptation occurs after phage infection ( 489 ) ; even 
ultraviolet-irradiated phage, which is unable to reproduce but can still 
kill the bacteria, prevents enzymatic adaptation. It is well to remember 
that enzymatic adaptation requires actual synthesis of specific enzyme 

5. Only one enzyme activity is found to increase in E. coli B in- 
fected with phage T2. Desoxyribonuclease activity increases many 
times (508a). This may reflect an actual increase in amount of enzyme 
or only a decrease in a desoxyribonuclease inhibitor (383a). The des- 
oxyribonuclease may play a role in the breakdown of host DNA and 
possibly in the synthesis of phage DNA. 

The role of phage in the synthesis of new phage. The infecting 
phage appears to assume the direction of specific syntheses; it allows 
the bacterial machinery to produce the building blocks, but decides 
how these are to be put together. The mechanism by which the actual 
"patterning" of new protoplasm by the phage takes place remains un- 
known. It is on the same level as the mechanism of replication of 
any "self-reproducing" molecular structure in any cell. 

The phage-infected bacterium, however, must synthesize certain 
substances that were not present in the uninfected cells. In fact, the 

180 Bacteriophage-Bacterium Interaction 

DNA of the T2, T4, T6 phages contains, instead of cytosine, the 
5-hydroxymethyl-cytosine (687b). This compound, as far as we know, 
is not present either in the bacterial host or in any other biological 
material. The synthesis of phage DNA must entail a synthesis of this 
pyrimidine. Thus, the infecting phage does not simply redirect the 
assembly of building blocks produced by bacterial enzymes, but ac- 
tually determines the synthesis of specific compounds.* This determi- 
nation might or might not require the intervention of enzymes con- 
tributed by the phage. It is possible, for example, that the presence 
in the cells of the vegetative phage primes or stabilizes the synthesis 
of the new compound by bacterial enzymes. 

Phage infection and bacterial organization. A suggestion of the 
nature of the process by which the phage replaces the pattern of 
bacterial protoplasm with its own is provided by cytological observa- 
tions (445; 498). One of the first changes noticeable in a bacterium 
after phage infection (within 2-3 minutes) is an alteration of the 
morphology of the bacterial nuclei, the DNA-containing bodies. In 
fixed and stained preparations of bacteria infected with phages like 
T2, the nuclei are seen to be broken into chromatic blocks, which move 
to the periphery of the cell. If the infecting phage is active, the nu- 
clear disruption is followed several minutes later by appearance of new, 
granular chromatin, which is probably phage. Infection with inactive 
but killing phage (inactivated by ultraviolet) causes nuclear disrup- 
tion and simply leaves the cells without organized DNA. With phage 
T5, where the latent period is longer, there is an interval of time be- 
tween nuclear breakdown and appearance of new chromatin. During 
this interval no cell structure giving the cytochemical reactions of DNA 
is recognizable, although DNA is certainly present in the cells (fig- 
ure 62). 

The nuclear changes provide an explanation of the killing of bacteria 
by phage that cannot reproduce. They also suggest that the redirec- 
tion of syntheses may be due to the elimination of the bacterial genes 
as determiners of protoplasmic specificity, with their replacement by 
phage genes in that role. 

As for the synthesis of specific viral components, such as hydroxy- 
methyl-cytosine, by phage-infected bacteria, the cytological picture 
suggests that the role of the phage may be to introduce specific 
genetic determinants for the production of the needed enzymes. 
Thus, the virus would indeed introduce new enzyme systems into the 
host cell, not, however, as preformed phage enzymes, but as genetic 
patterns for enzyme synthesis. 

Figure 62. Cytological changes in Eschericliia colt, strain B infected with phage 
T5. a-d, Chromatin stain (HCl-Giemsa). (a) Uninfected cells; (b) 15 min- 
utes after infection; the infected cells show no organized chromatin; (c) 30 min- 
utes after infection; the infected cells show variable amounts of granular chro- 
matin; (d) 48 minutes after infection; the infected cells are filled with chromatin 
and undergo lysis. eh t Cytoplasmic stain (thionine). The chromatin bodies re- 
main unstained, (e) Uninfected; (/) 10 minutes; (g) 30 minutes; (h) 45 min- 
utes: Note that this method gives the reverse appearance of the chromatin stain. 
(Courtesy Dr. R. G. E. Murray, University of Western Ontario, London, Ontario. 


182 Bacteriophage-Bacterium Interaction 

An interesting confirmation of this viewpoint is an observation on 
coli-cells sterilized by nitrogen mustard. These cells synthesize no 
DNA but accumulate RNA. Infection with phage T2 causes an im- 
mediate stoppage of RNA synthesis and a resumption of DNA syn- 
thesis, followed by lysis with phage liberation. Here the phage seems 
to have contributed the necessary DNA pattern for phage synthesis 

The postulated substitution of phage genes for host genes as a result 
of infection has been described as "parasitism at the genetic level" 
(437). We shall see in chapter 9 that the situations observed with 
lysogenic bacteria fit, and in many ways enlarge, the concept of an 
intimate tie-up between the genetic apparatus of the host and that of 
the virus. 

The cytochemical events observed with the T phages are not present 
in all phage-host systems, however. With many other systems, RNA 
and enzyme syntheses continue until lysis, and the only observable 
change in synthetic pattern is a delay of several minutes in DNA syn- 
thesis (603a). This is true of temperate phage, which can establish 
lysogeny, and also of their virulent mutants. This suggests that the 
shift in production from host DNA to phage DNA is the critical step, 
whereas the other changes in synthetic patterns may be restricted to 
special cases. % 



The Bacteriophage-Bacterium System 


Examination of the phage produced by infected bacteria and of its 
relations to the infecting phage provides a great deal of information 
on phage development. The amount and composition of the new 
phage as a function of the amount and composition of the infecting 
phage permits us to analyze the genetic relations between the two. We 
shall discuss, first, the effects resulting from intracellular phage changes, 
and then the effects that follow modifications of the phage input 

Spontaneous phage mutations and the rate of phage reproduction. 
Bacteriophages as well as other viruses undergo spontaneous mutations 
(see chapter 15). Among the phage mutants, the most useful for 
genetic studies belong in two categories. The "host-range" mutants 
(h) are recognized by their ability to grow on bacterial strains re- 
sistant to the normal strain of phage (434). Thus phage T2/i is active 
not only on E. coli y strain B, but also on strain B/2, which is a mutant 
of E. coli B resistant to T2. An h mutant attacks one or more strains 
of bacteria that are resistant to the parent phage. Sometimes the 
resistant bacteria have acquired their resistance by mutation from a 
sensitive bacterium; sometimes they are found to be resistant upon first 
isolation. The h mutants are easily recognized by plating with the 
resistant bacterium, on which they form plaques; the original, wild- 
type phage does not. 

The plaque-type mutants of phages T2, T4, T6 belong to a series 
(see figure 63) characterized by plaques that are either smaller or 
sharper or larger or more completely lysed than those of the wild 

1 The student should be familiar with the basic facts of genetics to understand 
the material discussed in the following sections. 



Bacteriophage-Bacterium Interaction 

type (309). Most of these variations reflect quantitative changes in a 
property of these phages, the so-called "lysis inhibition": if a bacterium 
already infected is reinfected with phage before lysis occurs, it lyses 
late and produces a higher yield of phage (181). Mutations of the 
plaque-type series lead either to elimination of lysis inhibition (mu- 
tants r = rapid lysis) or to a reduced inhibition (mutants w = weakly 
inhibited). Exaggerated inhibition may be responsible for the ra type 

Figure 63. Plaques of phage T4 and of various plaque-type mutants. 1, wild 
type; 2, m (minute); 3, tu (turbid); 4, tu tu (double turbid mutant); 5, r (rapid 
lysis); 6, r tu; 7, r m; 8, r tu tu. Note the suppressive action of non-r plaques on 
the development of neighboring r plaques. Courtesy Dr. A. H. Doermann, Oak 
Ridge National Laboratory. 

(minute plaque). The various mutant plaque types are easily recog- 
nized; they are frequently observed in platings of lysates of the T-even 
phages, the frequency of mutants being often between 0.01 and 1%. 

The mutations that give rise to these phage mutants take place only 
during intracellular phage reproduction and are apparently spon- 
taneous events, in the sense that the factors that control their occur- 
rence are unknown. We only know that these factors act in a com- 
pletely random fashion. 

Let us now consider the mutation process more closely. When a 
mutation occurs in a bacterium, the phage yield contains both normal 
and mutant phage (434). The amount of the mutant phage present 
in the yield of the bacterium in which the mutation has occurred must 
depend on the mode of phage reproduction, or at least on the mode of 

CH. 9 Phage Mutations and Reproduction 185 

reproduction of the portion of genetic material ("gene") responsible for 
the determination of the mutated character. New phage genes must 
arise by a replication of some sort from preexistent genes. If^a muta- 
tion has a constant probability of occurring at each replication, the 
number of mutants to which a mutation gives rise in the bacterium 
depends on the sequence of replication. For example, if each phage 
replica is produced independently of the other replicas, mutated 
replicas will be distributed at random among the infected bacteria. 


Figure 64. Diagrammatic representation of possible sequences of production 
of \irus elements. A, Increase by repeated reduplications. B, Increase by suc- 
cessive replications of the initial element. C, Increase by replication of the last 
(dement produced. Solid dots indicate mutants. Mutants are produced in clones 
of identical sibs in case A, singly in case B, and in series in case C. Case A agrees 
with the experimental findings ( Luna, 438). From: Luna, Ann. Missouri Botan. 
Garden 32:235, 1945. 

If a phage replica, on the other hand, can in turn reproduce and give 
rise to further replicas, and so on, the mutated replicas will often be 
in groups or "clones" of mutated sibs.- These possibilities are illus- 
trated in figure 64. 

By counting all the mutant plaques T2r in the individual yields of 
several thousand bacteria infected with phage T2, it is possible to 
show that the mutants, when present in a bacterial yield, are grouped 
in clones of identical sibs (438). The distribution of clones of various 
sizes (1 mutant, 2 mutant, 3, 4, 5 mutants per clone) is very close 
to a distribution that is calculated by assuming that the reproduction 
of phage genes is an exponential process, following the equation 

N = 2 n [13] 

where N is the final number of gene copies and n corresponds to the 
number of phage generations that have taken place. 

2 A "clone" is the group of all the indi\ iduals derived from one ancestor by 
vegetative reproduction. 

186 Bacteriophage-Bacterium Interaction 

This equation is formally analogous to the equation for clonal repro- 
duction of bacteria, protozoa, or any other vegetative cell population. 
It tells us that the phage genes reproduce from 1 to 2 to 4 to 8, rather 
than by the successive production, 1 by 1, of single copies of the origi- 
nal gene brought into the bacterium by the infecting phage particle. 

This result, we must remember, concerns individual phage proper- 
ties; in itself it does not necessarily mean that all phage genes reproduce 
together as a single particle, giving rise to 2 particles, which then 
produce 4, and so on. All it means is that individual determinants of 
phage heredity reproduce exponentially by successive duplications. 
Whether or not they do so all together, assembled in a single element, 
has to be shown by other methods. 

Multiple and mixed infection. When several phage particles attack 
the same bacterium, we speak of multiple infection if they are all alike, 
of mixed infection if they are of different types. The results are dif- 
ferent and depend on the degree of similarity between the particles. 
Multiple infection gives results quite similar to single infection. The 
latent period, the phage yield, and the time of appearance of mature 
particles, as shown by premature lysis, are not significantly different. 
This similarity between single and multiple infection is not due to a 
limitation in the number of infecting particles that can penetrate the 
cell; rather, it-indicates that the amount of phage produced in a bac- 
terium is limited by some other mechanism. 

When multiple infection consists of two infections with the same 
phage, separated by an interval of several minutes, the second phage 
contingent, after being adsorbed, is often actually destroyed, and part 
of its phosphorus is shed into the medium in soluble form (238). This 
may mean that the phage skin and nucleic acid separate, and that the 
nucleic acid is partly released outside. This second phage contingent 
does not participate in reproduction, nor does it contribute phosphorus 
to the phage progeny (237). It may, however, give sign of its presence, 
for example, by lysis inhibition. 

Mixed infection with unrelated phages. Experiments on mixed in- 
fection are done using indicator strains of bacteria, which selectively 
reveal each phage in a mixture of phages. The usual indicator strains 
are bacterial mutants resistant to one or another of the phages. The 
spontaneous mutations of bacteria to phage resistance have proved very 
useful in the study of bacterial genetics (442). A mutation may pro- 
duce resistance to one or more phages (see table 22). There is no 
obvious relation between the degree of relatedness of two phages and 
the cross-resistance pattern of various bacterial mutants to them (177). 

CH. 9 

Mixed Infection 


Table 22. Bacteriophage resistant mutants of E. colt, strain B, with 
respect to bacteriophages T1-T7 

Modified from Luria G}#>). 

Each mutant is derived from wild type by a single mutational step. 





r ^ 



T2 T6 T4 

T3 T7 



Wild type 



S S S 

S S 

Frequent mutants 




S S S 

S S 


B/l, 5 



S S S 

S S 




S R S 

S S 

B/3, 4 



S S R 

R S 


B/3, 4, 7 



S S R 

R R 

B/3, 4, 7 



(S) (S) R 

R R 

Complex mutants 

B/l, 3, 4, 7 



(S) (S) R 

R R 


B/l, 3, 4, .>, 7 



(S) (S) R 

R R 

B/*, 3, 4, 6, 7 



R R R 

R R 

B/l, 2, 3, 4, 0, 7 



R R R 

R R 


* The braces indicate serological relationship between phages. 
( ) The parentheses indicate ability to produce phenotypically modified phage 
(see page 197). 

S = sensitive; R = resistant. 

Given a pair of phages, it is almost always possible to obtain a bacterial 
strain sensitive to one phage and resistant to the other. These resistant 
mutants are used as indicators to identify the phages in a mixture, in 
the following way ( figure 65 ) . 

Suppose we plate a mixed suspension of phages Tl and T2 with 
bacteria of the mutant strain B/l, resistant to Tl and sensitive to T2; 
only phage T2 will produce plaques. If we use strain B/2 instead of 
B/l, only Tl will produce plaques. If we plate our phage suspension 
with mixed indicators, that is, with a mixture of bacteria B/2 and 
B/l, all plaques will be "turbid," since the plaques formed by each 
phage by lysis of the sensitive bacteria will be overgrown by the other 
bacterial type resistant to that phage (168). Clear areas will be ob- 
served where plaques of the two phages overlap. 

This technique also makes it possible to test whether a bacterium 
sensitive to two phages, when infected by both of them, liberates both 


Bacteriophage-Bacterium Interaction 

types. The host bacteria are infected with the two phages, and a series 
of samples are plated before lysis, with single and with mixed indi- 
cators. A bacterium that liberates only T2 gives a plaque on B/l, no 
plaque on B/2, and a turbid plaque on the mixed indicators. A bac- 
terium that liberates only Tl gives a plaque on B/2, none on B/l, and 
a turbid plaque on mixed indicators. Only bacteria that liberate both 
Tl and T2 give clear plaques on the mixed indicators. 

(b) (c) 

Figure 65. Platings of a mixture of bacteriophages Tl and T2 on various indi- 
cator bacteria, (a) Indicator B/l, sensitive to phage T2 only, (b) Indicator 
B/2, sensitive to phage Tl only, (c) Mixed indicators B/l and B/2; both phages 
form turbid plaques, with clear areas of lysis where the plaques overlap. En- 
larged about 

The results of mixed infection depend on the degree of relatedness 
of the infecting phages. In the T series, if the phages are unrelated 
by morphological and serological criteria, the resulting phenomena, 
which have been grouped under the general term of interference ( see 
chapter 14), can be summarized as follows (168): 

1. There is no interference in the adsorption process. Adsorption is 
normal for both phage types, until extensive coating of the bacterial 
surface becomes a limiting factor. 

2. There is complete mutual exclusion. A given bacterium liberates 
either one or the other of the unrelated phage types, never both. 

3. There is a depressor effect: the excluded phage causes a reduction 
in the yield of the successful one. This depressor effect can be counter- 
acted, within several minutes after infection, by addition of antiserum 
against the excluded phage. This suggests that the depressor effect is 
exerted by some portions of the excluded phage that are located on 

CH. 9 Genetic Recombination 189 

the bacterial surface and are still accessible to antiserum, possibly the 
skins of the excluded phage. 

The mechanism of mutual exclusion is not yet clear. As we shall 
see later (page 202) a phage coming from outside may sometimes ex- 
clude maturation of a phage already present inside the bacterium, that 
is, of a lysogenically carried phage (669). In these and probably in 
all cases, mutual exclusion may involve a competition for some single 
piece of machinery needed for phage maturation. It is possible that 
the excluded phage multiplies in the mixed-infected bacteria and 
simply fails to mature. 

Thus, mutual exclusion and depressor effect may result from different 
types of interactions in separate regions of the infected cell. The most 
remarkable feature of mutual exclusion between unrelated phages is its 
all-or-none character, no exception to which has been reported. 

Mixed infection with related phages. When serologically related 
phages such as T2, T4, and T6 infect the same bacteria, mutual ex- 
clusion fails and most bacteria liberate particles of both types (173). 
If the phages are very similar, for example, if they are mutants of the 
same phage, every bacterium is a mixed yielder. In a mass lysate 
derived by mixed infection with a pair of mutants of the same phage, 
the ratio of particles resembling the two parent types is generally the 
same in the yield as in the infecting mixture. This indicates that little 
intracellular selection takes place for or against either phage type. 

Participation of many infecting phage particles in reproduction is 
only possible, however, if infection is almost simultaneous. A few 
minutes of delay cause an exclusion between related phages (195). 
This exclusion is accompanied by destruction and partial release into 
the medium of the excluded particles (238). 

Genetic recombination. When two related phages, differing from 
one another in at least two characters, infect the same bacteria, a new 
phenomenon is observed. This consists of the appearance of phage 
types that were not present in the input and that represent new com- 
binations of the characters of the parents. Thus, infection with T4r 
and T2 gives rise to four types: T2, T2r, T4, T4r; infection with T2h 
and T2r gives T2/i, T2r, T2/ir, and T2 (173; 310). 

The discovery of this "recombination" opened a new field in virus 
genetics. It soon became clear (314; 315) that each character that 
acted as a recombination unit represented one of two alternative forms 
of some discrete phage portion or gene, and that recombination was a 
reshuffling of these genes. Thus, we should write more correctly and 
in complete analogy with usual genetic terminology: 

190 Bacteriophage-Bacterium Interaction 

T2/ir+ X T2h + r -> T2fcr+, TA+r, T2h 

The superscript + indicates the normal (wild-type) form or allele. 
A single-step mixed-infection experiment of this kind is called a cross. 

The important discovery was soon made (314) that independently 
isolated r mutants of the same phage are seldom, if ever, identical, and 
are actually not allelic, since upon mixed infection with two r mutants 
one obtains some r+ recombinants: 

X T2r 2 - Tar^r), T2r 2 (r), 

The letters in parentheses indicate the plaque character or phenotype 
of each phage. The scheme of figure 66 gives the genetic interpreta- 
tion of these crosses. 

It is clear that phenomena such as this may reveal essential features 
of the process of phage development; this process must be such as to 
allow for recombination of phage characters to take place. A quanti- 
tative analysis of genetic recombination may be as informative on 
phage development as the quantitative analysis of genetic data in 
higher organisms is in clarifying the nature and organization of the 
genetic material in their germ cells. A peculiarity of phage genetics 
is that quantitative data on recombination, that is, on the relative fre- 
quencies of various classes of progeny from a. cross, are obtained by 
examining the yield of many bacteria or, at best, the whole yield of 
individual bacteria. The exact number of phage particles of each 
parental type that infect individual bacteria is unknown; all we know 
is the average number of the infecting particles. Moreover, the process 
of recombination may occur many times within each bacterium. Thus, 
the final result we observe reflects an unknown number of crosses 
among an unknown number of parental particles and of their offspring. 
It has been pointed out that mixed infection of a single bacterium is 
equivalent to an experiment in population genetics rather than to a 
cross between two individuals. 

The most salient features of the phenomenon of genetic recombina- 
tion, as studied mainly with phages T2 and T4, may be listed as 
follows (183; 184; 314; 315; 657): 

I. In mass populations derived in single-step experiments from bac- 

teria mixed-infected with phage of two types: 

1. The yields of the two recombinant types are approximately 
equal and characteristic for each pair of character differences 
studied; for example, a cross T2/i X T2r yields equal numbers 
of T2hr andT2h+r+. 

CH. 9 

Genetic Recombination 


2. Some characters are recombined independently (unlinked 
genes). In a mass lysate the frequency of each recombinant 
type in these cases is generally lower than 25% of the total, 
the proportion expected if there were random assortment of 


I. Cross: T2xT2*r 2 . Progeny] to^TSA 

Interpretation of recombinant progeny 






II. Cross: T2r t x T2h. 

Interpretation of recombinant progeny 
^ H II h 





Recombinant Recombinant 
T2 T2hr 

Linkage map of phage T2 

Figure 66. 

"minute plaque" locus. 
"ultraviolet sensitivity level" locus. 

The chromosomal interpretation of genetic recombination in bac- 

two independent pairs of elements. The limitation is inter- 
preted as due to the fact that not all the progeny particles have 
had an equal chance for biparental origin. The more phage is 
produced in a bacterium (for example, under conditions of 
lysis inhibition) the closer the recombinant frequencies be- 
come to the expected 25% value. In a cross between parents 
differing by 3 instead of 2 unlinked characters, the recombi- 

192 Bacteriophage-Bacterium Interaction 

nants for a pair of characters are divided into two equal 
classes, each of which carries one of the two forms of the third 
character. For example, in a cross T2rhm X T2r f h+ra f (r, 
h, m unlinked) one-half of the rh+ (or of the r * h) is w, the 
other half is ra + . Only three unlinked groups of characters 
have as yet been described in phage T2 (see figure 66). 

3. Linked characters, which do not assort independently in a 
cross, have been found. They give frequencies of recombi- 
nants intermediate between the minimum observable (about 
1%) and 25%. The results are compatible with the assumption 
that the corresponding genes lie arranged linearly on "chromo- 
somes" and that the recombination frequencies reflect "cross- 
over frequencies" or "distances" between genes on the same 
chromosome, that is, additive probabilities that groups of 
linearly arranged genes become separated by crossing over. 
Thus, a "3-point test" (using 3 linked characters a, b, c) tells 
whether a certain gene lies between two others or beyond one 
of them. If c is between a and &, the frequency of recombi- 
nation for the pair a, b is approximately the sum of the fre- 
quencies of recombination for a, c and for b, c. The linkage 
map of figure 66 was constructed after experiments of this 
type. The validity of the linkage hypothesis is corroborated 
by "coupling and repulsion" experiments; if a cross ab X AB 
gives, say, 3% aB (and Ab), the cross aB X Ab gives 5% AB 
(and ab). For linked characters as well, the higher the phage 
yield, the higher the proportion of recombinants becomes. 

4. A recombinant particle may exhibit characters derived from at 
least 3 parents. Thus, in a cross abC X aBc X Abe, some 
particles ABC are formed. These could not be formed by any 
type of exchange between two parental particles only. If, 
therefore, the elementary act of recombination is a mating of 
phage particles in pairs, followed by redistribution of genetic 
elements, there must then be several rounds of matings in 
each bacterium. 

5. Recombinants are already present among the first mature 
phage particles that are obtained by premature lysis at the 
earliest time after infection; the frequency of recombinants in 
the premature yield is lower than in the final yield. 

II. In single-burst experiments from mixed-infected bacteria: 

6. The numbers of recombinants of opposite type produced in 
individual bacteria are neither identical, nor similar, nor even 

CH. 9 Heterozygotes 193 

significantly correlated. This holds both for final and for pre- 
mature yield. If recombination is due to mating in pairs fol- 
lowed by reciprocal exchanges, the result of the individual 
pairings should be the production of equal numbers of the 
two recombinants. Possibly, some other event occurring later 
may distort the numbers of opposite new types that the mat- 
ing produces. Unequal reproduction of recombinants after 
the exchanges could do so. Such unequal reproduction may 
reflect partly the nonsynchronization of phage multiplication, 
partly the removal of particles from the reproducing popula- 
tion by maturation. 

As a whole, the recombination experiments reveal a tremendously 
complex genetic system in phage. Genetic recombination has been 
observed in all the coli-phages of the T systems in which it has been 
looked for (Tl, T2, T3, T4, T5, T6). How it takes place we do not 
know with certainty. The most likely explanation seems to be a series 
of matings by random pairings, irrespective of genetic type, among 
haploid, 3 vegetative phage particles that have not yet become mature 
and infectious. Recombination in the paired phage would occur both 
by random segregation of chromosomes and by "crossing over," as in 
sexually reproducing higher organisms. Indeed, it is possible to pre- 
dict the actual frequencies of recombinants in various crosses (for 
example, with parents in unequal numbers ) and the lack of correlation 
between reciprocal recombinants on the basis of a model (657) con- 
sisting of repeated, completely random matings in a population of 
reproducing vegetative phages, with maturation (see page 170) re- 
moving individual phages at random from the sequence of reproduction 
and recombination. The remarkable feature is, of course, the tremen- 
dous frequency of repeated matings in the short period of a single 
intracellular cycle. 

Heterozygotes. If matings among vegetative phages do occur, it 
would be desirable to demonstrate the existence of mating forms. 
Among the phage particles comprising the yield of any one cross, about 
2% give progeny that contain particles of two types (312). For 
example, in the yield from a cross of T2r X T2r + , about 2% of the 
new particles give' progeny consisting of r and r+ particles in equal 
amounts. These "segregating" particles produced in the cross are, 
then, heterozygous for the r locus. The heterozygotes, although they 

3 "Haploid*' indicates a nucleus containing a single set of chromosomes; "diploid," 
a nucleus with two sets of homologous chromosomes. 

194 Bacteriophage-Bacterium Interaction 

fit the general picture of a mating mechanism, are difficult to explain 
in terms of the mating theory discussed above, especially because the 
heterozygous portions of their chromosomes appear always to be very 

A study of linkage in heterozygotes of T2 (416a) has shown that 
the heterozygous regions of the chromosomes are those at which cross- 
overs have occurred. The results suggest a special mode of phage 
reproduction, in which the chromosomes are replicated by a linear, 
zipper-like mechanism. Partial replicas, initiated at opposite ends of 
two homologous chromosomes, could join together and give rise to 
chromosomes with overlapping regions. If any allelic differences occur 
in the duplicated, overlapping regions the particles will be hetero- 
zygous for the corresponding genetic properties. 


Radiation treatments have been most valuable in attempts to clarify 
the intracellular phage development by modifying phage in such a 
way that its development stops at one stage or another. Thus we learn 
that phage adsorption is not necessarily followed by death of the bac- 
teria, since x-ray-inactivated phage is adsorbed and disappears without 
stopping bacterial growth (665). Likewise, the following stage of in- 
vasion, characterized by bacterial death and nuclear disintegration, 
can be dissociated from production of new phage by ultraviolet treat- 
ment of the phage before adsorption ( 441 ) . Moreover, new phenom- 
ena are observed in bacteria infected with irradiated phage, depending 
on the conditions prevailing upon infection. 

Reactivation phenomena. An inactive phage particle is defined as 
one that fails to reproduce in the host bacterium. Upon adsorption of 
such an inactive particle (if adsorption can still occur) lysis fails to 
take place and even artificial disruption of the bacterium reveals no 
active phage. If bacteria infected with phage inactivated by ultra- 
violet light (UV) are exposed to visible light or near UV (3100- 
5000 A) a fraction of the bacteria produce active phage. This photo- 
reactivation (193), an instance of a very general occurrence with cells 
modified by UV, is characteristic for each phage, is barely noticeable 
after x-ray inactivation, and has not been observed with phage inacti- 
vated by chemicals. No photoreactivation occurs if the irradiated 
phage is exposed to light in the free, unadsorbed state. 

For phages like T3 or T7, the damage made by each UV quantum 
must be repaired by 1 light quantum, although some of the UV damage 

CH. 9 Radiation Analysis 195 

is irreparable. For phages like T2, the situation is more complicated, 
and the hypothesis of a 1 : 1 repair of individual damages requires some 
subsidiary assumptions (98a). The pigment responsible for the photo- 
chemical reaction may be a flavine. The dependence of photore- 
activation on factors such as temperature probably reflects the effect 
of these factors on the production and utilization of the pigment. 

Another type of reactivation is observed with several phages upon 
multiple infection with irradiated particles. Active phage is produced 
in many more bacteria than those which receive the few phage particles 
that register as active in single infection (443). This multiplicity 
reactivation indicates a cooperation among damaged phage particles 
in "pulling through" in the same bacterium. The particles must be 
either of the same type or of nonmutually exclusive types. It was 
thought at first that multiplicity reactivation was due to genetic re- 
combination, with the formation of active particles by transfers of 
genetic material among genetically damaged particles. Further work 
showed however that the genetic recombination hypothesis was inade- 
quate to fit all the aspects of the phenomenon (194). Multiplicity 
reactivation is probably due to a cooperation among damaged particles 
in performing some early step in the integration between phage and 
host cell. Reproduction can then occur if at least one of the particles 
is genetically undamaged. Such a genetically undamaged particle 
might arise by genetic recombination. 

Intracellular irradiation. By exposing phage-infected bacteria to 
radiation, their ability to produce phage can be suppressed (447). In 
this case radiation acts on the virus itself rather than on the bacterial 
host. Actually, most phages can grow in bacteria that before infec- 
tion had received doses of radiation much greater than those needed 
to suppress phage growth. This situation is not surprising, since we 
have already seen that the part of the bacterium used in phage syn- 
thesis is its enzymatic machinery (which is very resistant to radiation) 
rather than its more radiosensitive nuclear apparatus. 

The analysis of intracellular suppression of phage as a function of 
radiation dose gives different results with different phages. With T7, 
a small and probably simple phage, the suppression curve in single- 
infected bacteria immediately after infection is identical with the in- 
activation curve of free phage. Then, after several minutes, it changes 
into a "multiple-hit" type of curve. The final slope, that is, the proba- 
bility of an individual hit, remains approximately constant (64; figure 
67). This indicates that suppression of the ability of an infected bac- 
terium to liberate phage requires the inactivation of all the intracellular 

196 Bacteriophage-Bacterium Interaction 

virus elements, and that the individual elements to be inactivated retain 
the same radiosensitivity as the free virus particles, while their number 

The situation with T2 is more complex (64; 447). The radiosensi- 
tivity of the infecting particle diminishes for several minutes, before 

100 200 

UV dose in seconds 


Figure 67. The intracellular inactivation of phage T7 by ultraviolet ligbl 
Abscissa: dose of UV ligbt. Ordinate: tbe proportion of infected bacteria whic! 
retain the ability to liberate phage. Bacteria, infected at time 0, are chilled afte 
various times of incubation, irradiated with UV, then plated for assay of plaque 
forming ability. The number near each curve gives the time in minutes bctwee 
infection and chilling. From: Benzer (64). 

any increase in the number of radiosensitive units is seen. This be 
havior fits the idea that the phage has to perform an initial reactior 
which requires the intervention of a specific radiosensitive portioi 
This portion may be needed only in the first few minutes, after whic 
the phage reproduces in the form of less radiosensitive units. Thes< 
late in the latent period, give rise to mature particles. 

CH. 9 Changes Induced in Phage by the Host 197 


From all lines of evidence, we can now construct a tentative picture 
of the events of phage development. This picture is, of course, pro- 
visional, and serves mainly a purpose of orientation. 

With a phage like T2, irreversible adsorption is followed by a process 
of invasion. This depends on a penetration of the nucleic acid-con- 
taining phage portion into the bacterium. It is characterized, on the 
one hand, by a modification of the cell surface, in which a mechanism 
for the rejection of additional phage particles becomes established; 
on the other hand, by an incapacitation of the nuclear apparatus of the 
bacterium. In these stages, the phage is nonrecoverable in active 
form (vegetative phage). A radiosensitive phage portion becomes 
unnecessary and is possibly used up. Then, logarithmic reproduction 
of the genetic phage elements begins, accompanied by genetic recom- 
bination. The formation of infectious phage particles, which starts 
about the middle of the minimum latent period, is preceded by the 
appearance of immature elements related to various portions of the 
phage particles. The maturation probably consists of the acquisition 
of a "cytoplasm" as opposed to a "nucleoplasm" or genetic core of the 
phage. The cytoplasm includes the protein skin, the tail, and possibly 
other parts of the mature particle. Maturation renders a phage particle 
infectious for other bacterial cells and removes it from further partici- 
pation in reproduction in the cell in which it was formed. 

Other phages may be less differentiated into portions of different dis- 
pensability and of different function; they also differ in the extent to 
which they disrupt host functions and organization. 


The concept of a physiologically active, nongenetic portion of the 
phage, acquired late in phage development, is supported by the fact 
that growth in different hosts can induce transitory changes in phage 
particles (76; 446). An example of such phenomena, which are ap- 
parently fairly frequent and will be discussed in greater detail in 
chapter 15, is illustrated in figure 68. A phage, for example P2, when 
grown on a host strain Sh, is generally unable to attack another host B. 
A few particles attack B, however, and give rise to a new phage form 
(P2B) that can grow regularly on B. The change is not permanent, 


Bacteriophage-Bacterium Interaction 

however, since a single cycle of growth of P2B on Sh returns it to the 
P2 form, which does not grow on B. 

Changes of this type are strictly host controlled. They modify the 
ability of a virus to grow on a given host. In some cases, as in the 




Figure 68. Scheme of host-induced modification. The percentage figures near 
the arrows correspond to the fraction of cells in which the phagc can grow. Phage 
P2, after 1 cycle of growth in strain Sh, can only grow in 1 of 10 4 cells of strain B. 
After 1 cycle of growth in strain B, P2 is in the P2B form, which can .grow in every 
cell of either B or Sh. Modified from: Bertani and Weigle (76). 

one mentioned above, they allow repeated growth on the new host. 
In other cases, instead, they may prevent continuous growth on it. 
The importance of such phenomena is that they represent the only 
known modifications of virus properties directly induced by the host. 
The modifications are not produced by selection of spontaneous phage 
variants; they are temporary phage changes due to the intracellular 
environment in a particular host. 

CH. 9 Lysis of Bacteria 199 

How the host does it, we do not know. It may provide the phage 
with some special substrate, or it may force the phage to develop 
differently because of the lack of some substrate, or it may contribute 
some specific structural patterns to the virus. 


The mechanism by which phage causes bacterial lysis is unknown. 
Lysis is not tied to production of infectious phage, since it may take 
place in the presence of proflavine, which allows formation of incom- 
plete phage particles (175). Phage liberation has been claimed to 
occur in some cases before visible lysis, possibly as a result of a local- 
ized surface damage to the bacterial membranes, which may allow the 
cell contents to escape. Usually, lysis accompanies phage liberation 
and appears to be a rather sudden breakage of the cell membranes, 
which often leaves an empty cell wall that later slowly disintegrates 
(see figures 30-31). 

Nothing is known as to the chemical or physicochemical mechanisms 
of lysis; these may involve the action of bacterial enzymes, of phage 
enzymes, or of both. On the one hand, the occurrence of bacteriolytic 
substances produced by bacteria under stimuli other than phage, and 
also upon lysis by phage (270), seems to favor a role of bacterial 
enzymes in the lytic process. On the other hand, the specific duration 
of the latent period between infection and lysis, which for the same 
bacterium may vary from 13 to 50 minutes or longer depending on the 
phage, suggests a role of phage in determining the course of lysis. The 
latent period remains constant even when the bacterial generation time 
is varied five-fold, by changes in medium composition. This indicates 
its independence of the rate of bacterial syntheses. In mixed infection, 
with a long-latent-period phage like T2 excluding a short-latent-period 
phage like Tl, lysis and liberation of T2 occur after the latent period 
characteristic of the excluding phage. 

Bacteriolytic substances have been obtained from heavily irradiated 
or osmotically disrupted bacteriophage particles, but their role in the 
normal process of bacterial lysis is unknown. In some cases the lytic 
action is due to the isolated skin or "ghost" of a disrupted phage par- 
ticle (308). Upon lysis by phage, bacteria often release lytic sub- 
stances acting upon other strains, but there is little evidence that these 
"lysines" are phage specific (270). 


Bacteriophage-Bacterium Interaction 


The prophage and its maturation. Cell-free filtrates of bacterial 
cultures often contain phages that lyse other strains of the same bac- 
terial species or of related species. Such phage-bacterium associations 
sometimes represent accidental associations of a phage with a resistant 

Figure 69. Centered plaques produced by plating a few washed cells of a lyso- 
genic strain with an excess of a sensitive strain. The sensitive bacteria are lysed 
by the phage liberated by the lysogenic colonies. 

bacterium that throws off enough sensitive mutants to maintain phage 
propagation in mass cultures; it is then possible to free the bacteria 
from phage by single-cell isolations. In a majority of the phage- 
bacterium associations, however, this is impossible; every cell is a 
potential source of phage (true lysogenesis). Plating washed lyso- 
genic cells with a layer of phage-sensitive bacteria (indicator bacteria) 
leads to formation of plaques, each plaque being centered by a colony 
of lysogenic bacteria (figure 69). In platings of unwashed whole cul- 
tures of lysogenic bacteria we also observe noncentered plaques, which 
originate from free phage particles present in the lysogenic culture. 

CH. 9 Lysogenic Bacteria 201 

The phages that establish lysogenic relations with their hosts are called 

There is no doubt that each cell of a lysogenic strain is a phage 
carrier (453). Phage liberation occurs by occasional lysis of a cell. 
The frequency of such spontaneous phage liberation was found to be 
1:200 cells per generation in one case, 1:50,000 in another. When 
frequent, the lysis of individual cells can actually be observed under 
the microscope and is found to be accompanied by liberation of phage, 
which occurs only at this time. A lysing bacterium liberates an amount 
of phage of the same order as the yield from a sensitive bacterium lysed 
by the same phage. Sometimes the lysogenic bacteria cannot adsorb 
the phage they carry; in most cases they can. They are generally not 
lysed by such reinfection with their own phage, to which they are, 
therefore, immune. 

Inside the lysogenic bacteria the temperate phage is as undetectable 
as a virulent phage is in the early part of the latent period in the lytic 
cycle. No infectious phage can be recovered by artificial breaking of 
the lysogenic cells. In some lysogenic bacteria, mass lysis with phage 
liberation can be produced artificially, after a latent period of 30-60 
minutes, by such means as irradiation, treatment with certain sulf- 
hydryl compounds, nitrogen mustard, H 2 O 2 , and ion-chelating agents 
(452; 454). This process is called induction. Even in the "induced" 
bacteria, premature lysis or breakage reveals infectious phage in the 
last third of the latent period only. 

Thus we have reason to believe that phage is present in the cell in 
a form, the so-called prophage, which is different from the infectious 
particles, and which before liberation must undergo a process of matu- 
ration. Inside the cell, the phage shares the fate and properties of 
all other cell constituents. For example, in spore-forming bacteria the 
prophage survives temperatures much higher than the free phage 
could, just as cell proteins in the spore withstand heat treatment much 
better than in vegetative cells (188). Apparently the prophage repro- 
duces without upsetting the cell economy, and only occasionally does 
a new situation arise which leads both to phage maturation and to cell 
lysis. Prophage and bacterial protoplasm are synthesized side by side; 
the maturation process, instead, is bound to events that preclude bac- 
terial survival. Even in bacteria in which lysis is induced by irradia- 
tion, the synthesis of bacterial components (RNA, enzymes) continues 
during the latent period before lysis. 

Is the interdependence between maturation and lysis valid for phage- 
bacterium relations in general? If so, the difference between the lyso- 

202 Bacteriophage-Bacterium Interaction 

genie and the lytic situations would reside only in the frequency with 
which the presence of phage sets in motion the unknown series of 
events leading, on the one hand, to phage maturation, on the other 
hand, to bacterial disintegration and lysis. 

Experiments on mutual exclusion between a prophage and a super- 
infecting, virulent phage are compatible with this point of view ( 669 ) . 
In lysogenic bacteria induced by UV, infection with a phage unrelated 
to the carried prophage can exclude the carried phage from the yield 
(i.e., can preclude its maturation). This exclusion can take place even 
when the virulent phage is added only a few minutes before the ap- 
pearance of mature particles of the carried phage. We are led to 
believe that induction does what the prophage is unable to do, that 
is, eliminates the functional integrity of the cell and allows the pro- 
phage to proceed toward maturation; this process can be stopped by 
infection with another phage. 

We may, thus, visualize a series of alternative rulers of cell destiny 
the bacterial nucleus, the prophage, and the exogenous phage. The 
prophage of a temperate phage cannot displace the bacterial pattern 
without help from inducing agents. In induced bacteria the latent 
prophage captures the lead, not immediately upon induction but after 
an interval, during which it can still be displaced by an exogenous 

The situation in multiple lysogenesis with unrelated phages fits this 
hypothesis. Each cell carries the various prophage types; but, when 
spontaneous phage liberation occurs, only one phage type matures in 
any one cell (74). Mutual exclusion of maturation among unrelated 
prophages may indicate that the conditions that allow the cellular con- 
trols to be overwhelmed permit a competition among prophage types 
for control of the maturation mechanism, always leading to total vic- 
tory of one of the competitors. Mutual exclusion may also indicate, 
however, that the factors that permit maturation act specifically on one 
or the other of the prophage types. Related prophages may mature 
together (359a). 

A prophage may prevent multiplication of some unrelated phages 
and not to others ( see table 23 ) . The mechanism of this interference 
is unknown. 

The condition of the prophage in the lysogenic cell. If there were 
only one prophage copy per cell, some mechanism should be present 
for its orderly reduplication and segregation along with other cell 
components. If the prophage were present in multiple copies, these 
could be assorted at cell division, like other nonnuclear components of 

CH. 9 The Condition of the Prophage 203 

the host cells. Nonlysogenic cells are exceedingly rare or altogether 
absent in lysogenic strains. Therefore, either the prophages are many 
or they are not assorted at random. The evidence given below seems 
to support the idea that prophage is present in single copy, at least in 
permanent lysogenicity, and may be transmitted by a mechanism con- 
nected with chromosomal reproduction. 

The ability of a phage to be carried lysogenically, that is, its tem- 
perate or virulent nature, is genetically controlled. A temperate phage, 
upon infecting a sensitive host, produces lysis or lysogenesis in com- 
parable proportions of cells. This relation explains why the temperate 
phages produce a type of plaque consisting of a partially lysed area 
covered with lysogenic growth. Virulent mutants can be obtained 
from temperate phages (100). They give clear plaques and induce no 

After infection of bacteria with a temperate phage, the cells that do 
not lyse may for several generations segregate into lysogenic and 
sensitive, noncarrier offspring (424). Later, no more sensitive off- 
spring are produced, and every cell derived from a lysogenic parent 
is lysogenic. 

Interesting results are obtained when these stably lysogenic bacteria 
are infected with related phages, which they generally adsorb (75). 
Some typical results are shown in table 23. Infection of a lysogenic 
bacterium with a temperate mutant of its prophage gives rise to a 
temporary mixed lysogenesis; for several cell generations the few cells 
that lyse give mixed yields. But after a few generations the prophages 
segregate, and ultimately either one or the other prophage remains in 
each cell. Double lysogeny occasionally results. 

Reinfection of a lysogenic bacterium with a virulent mutant of its 
prophage also leads to temporary mixed lysogenesis. This is followed, 
however, by return to the original condition of lysogenesis for the tem- 
perate phage. An exceptional finding of a return to sensitivity sug- 
gests that segregation of prophages has occurred, but that the virulent 
prophage, being inherently unable to establish permanent lysogenesis, 
becomes eliminated. Altogether, the segregation of related prophages 
and their mutual elimination are easily interpreted if we assume that 
permanent lysogenesis requires the fixation of a prophage to some site, 
which exists in single copy and which segregates in an orderly manner 
at cell division, for example, a specific chromosomal location. The 
"anchored" prophage can be removed or replaced, completely or in 
part, by another prophage of the same type. A weakly virulent mutant 
cannot establish itself as prophage, nor can it overcome the immunity 








S ^ 

.8 & 

2" -^ 

*S "2 







Bacteriophage-Bacterium Interaction 


+ S, 




^ 2 fe 




^5- ^ 


M " 1> 




S ^ o o t o ^o t o 




II coJ5 4, 

w ,j-- c fc Sc 

' c| - -2 


*C .5 



- ^ 2 S 2 fi E^-g'g. 

< <1 3^ oQ ^Q 55 r/3 ?/2 Q H-3* 


CH. 9 Relation of Prophage to Host Cell 205 

caused by a related prophage. Strongly virulent mutants, however, 
can overcome the immunity, reproduce, and mature as though the 
prophage were not present (see table 23). 

Genetic experiments on lysogenic bacteria agree with the above con- 
clusion (409). The sexual strain K-12 of Escherichia coli (412) may 
exist in either a lysogenic or a sensitive, noncarrier form with respect 
to phage lambda. In crosses of lysogenic by sensitive bacteria, lyso- 
genesis behaves as a single segregating factor, whose position in the 
chromosomal map of the bacterium can be located with proper linkage 
tests. This observation agrees with the idea of a chromosomal loca- 
tion of the prophage, although it does not exclude that what segregates 
is not the prophage itself but a genetic determinant responsible for 
ability to carry the prophage. 

Although the prophage may be intimately bound to the reproductive 
mechanism of the cell, it is not irreversibly bound. It can be lost 
either by substitution with related prophages, as shown in table 23, or 
occasionally by growing the bacteria under special nutritional condi- 
tions apparently specific for each case, such as low Ca f f concentration 
(135) or presence of glucose (202). 

If the prophage is carried in single copy in the host's nucleus, then 
the process leading to spontaneous or induced production of mature 
phage from a lysogenic bacterium must involve more than maturation. 
It must consist also of the transformation of the prophage into vege- 
tative phage and of the proliferation of the latter. Thus, the lysogenic 
and the lytic types of phage-host relation seem to have in common the 
full series of events (vegetative reproduction and maturation) that lead 
to the production of many infectious phage particles from an initial 
noninfectious element. 

The nature of the relation of the prophage to its host cell. The pro- 
duction of lysogenicity by controlled infection with temperate phages 
makes it unnecessary to consider the prophage of a lysogenic cell as a 
"normal" cell constituent that becomes a "new virus" upon accidental 
liberation. Yet the possibility of an intimate relation of the prophage 
with the reproductive system of the host bacterium suggests a degree 
of integration that goes beyond the metabolic level and reaches to the 
genetic constitution of the cell. Once a prophage is established in a 
cell, it behaves like a cell constituent. This constituent, however, may 
behave as a rebellious and destructive element and proceed to mature 
and to lyse the cell. If even more complete integration were to occur, 
resulting in a carrier state never followed by maturation, we might 
never suspect the presence of a prophage. 

206 Bacteriophage-Bacterium Interaction 

Observations on the genetics of bacteria of the genus Salmonella 
are of interest in this connection (695). Most strains of Salmonella 
typhimurium are lysogenic for one or more temperate phages that can 
infect other strains (99). Certain strains produce a filtrable agent FA 
(identical with a phage) that can transmit or transduce individual 
characteristics of the strain of origin to other strains, which are thereby 
stably changed and act in turn as sources of FA. The transducing 
agent, limited in activity range to salmonella cultures having in com- 
mon the somatic antigen XII, can transfer to other strains individual 
traits of the strain that has last liberated it. The transduced traits 
include nutritional properties, fermentative abilities, drug resistance, 
and specific antigens. Only one trait is transduced at a time. Pre- 
sumably, during maturation, the Salmonella phage may incorporate in 
its particle not only the phage nucleus but also some genetic elements 
of the host, possibly from the disrupted bacterial nucleus. These 
genetic elements are delivered into new host cells, which, if not lysed, 
may accept the new genetic elements in replacement of their own. 
This phenomenon brings the phage one step closer to a sort of sexual 
form of bacterial cell. 

There are other instances of host properties (other than phage sensi- 
tivity or resistance) that are changed upon acquisition of lysogenicity. 
In certain nontoxigenic strains of Corynebacterium diphtheriae the few 
cells that upon phage infection become lysogenic acquire the ability to 
produce the diphtherial exotoxin (236). These observations indicate 
a role of a lysogenically carried virus (or of something "transduced" 
with it) in the biochemical activities of the lysogenic cell. Nutritional 
differences occur between lysogenic and sensitive derivatives of other 
bacteria, for example, in Bacillus megatherium (164a). 

Lysogenicity, phage susceptibility and phage modification. A re- 
markable situation observed with some of the Vi-phages of Salmonella 
typhosa illustrates a role of lysogenically carried prophages not only 
in altering the susceptibility of bacteria to other phages but also in 
impressing upon these phages certain specific transformations. The 
original Vt-phage II, upon being plated with individual Vi-positive 
bacterial strains, gives a few plaques. These plaques yield "adapted" 
phages, which attack a common host strain A plus one or a few of 
the other Vi-positive strains (157 a). By this means, 31 adapted Vf- 
phages have been secured, by means of which 31 strains of Vf-positive 
typhoid bacilli can be "typed" ( 157, 213a). 

The adapted Vt-phages are not (or at least not all) host-range 
mutants of the original Vt-phage II. Upon a single transfer on the 

CH. 9 Coli-phages and Colicines 207 

common host strain A many of them revert to Vi-phage A, similar to 
the original, nonadapted Vi-phage II. The reverted Vt-phage A can 
then again be adapted by a single passage on a different Vf-strain of 
bacteria (13a). Thus, at least some of the adapted Vi-phages are not 
host-range mutants, but phenotypically modified phage variants (446; 
see pages 197 and 295). 

The discovery was made that the differences in Vt-type among many 
Vi-strains of S. typhosa are due to specific bacteriophages carried lyso- 
genically within them (13b). These latent, type-determining phages 
are completely distinct from the Vi-phages derived from Vf-phage II. 
The most remarkable element of this situation is that the latent pro- 
phages determine not only the host's susceptibility to unrelated phages 
but also the host's ability to modify phenotypically in a specific manner 
the Vt-phages that succeed in infecting it. The implications of these 
discoveries in relation to lysogenicity and host-controlled variation in 
viruses are not yet fully understood. The situation in Salmonella 
ttjphosa is certainly not unique, as shown by observations on lyso- 
genicity and phage susceptibility in other salmonellas ( 21 3b ) . It may 
indeed be an example of a widespread mechanism for control of virus 
susceptibility, not only in bacteria, but in other organisms as well. 

Coli-phages and colicines. Another observation of interest concerns 
the relation of lysogenic phage to the so-called colicines, substances 
produced by certain strains of E. coli and lethal for other strains (235). 
The colicines are not viruses, in the sense that they are not reproduced 
in the cell they kill. Yet most colicines are related, by similarity of 
host range, to specific coli-phages. Some colicine-producing E. coli 
strains undergo mass lysis, with production of colicine ( and no phage ) , 
upon treatment with the same agents that "induce" phage maturation 
in lysogenic strains (360). We get the impression that the colicine 
may be a nonreproductive maturation form of some lysogenically car- 
ried phages. They might be phage skins without nucleus or without 
the ability to release their nucleus into the host cell. 

The observations on temperate phages and on their lysogenic hosts 
have carried us far from the picture of a virus entering a host cell, 
reproducing in it, and becoming active and free again. Before our 
eyes the virus has become more and more a part of the host cell. The 
free, infectious form has been reduced to an incidental aspect of the 
life history of the phage. There is no denying the deep insight that 
these observations promise to give us into the nature pf host-virus rela- 
tionship and even into the genetic structure of the host. As we shall 
discuss in chapter 18, the facts of lysogenesis do not solve the question 

208 Bacteriophage-Bacterium Interaction 

of the origin of bacteriophage. The lysogenic condition may include 
examples of virus origin, or it may simply be the ultimate development 
of parasitism. 

Lysogenicity and phage ecology. The symbiotic relation between 
prophage and lysogenic bacteria, with occasional liberation, is much 
more favorable to the maintenance of the virus in nature than the lytic 
relation. The lytic relation is more frequently observed by superficial 
bbservation because of the dramatic occurrence of mass lysis, but is 
more likely to lead to virus extinction through rapid destruction of its 
hosts. Most phages probably exist in nature in some lysogenic bacterial 
reservoir, their occasional liberation providing them with a means for 
finding new hosts, which the phage may either lyse or render lysogenic. 
Virulent mutants will have little chances for survival, except under the 
artificial conditions of their selection and maintenance in the labora- 
tory. The lysogenic condition is extremely frequent. In certain bac- 
terial groups, for example, in Salmonella typhimurium, all strains are 
lysogenic for one or more phages that lyse other salmonellas (99). 
Indeed, one wonders if every bacterium may not be lysogenic, carrying 
phages most of which we have no means to recognize. It might even 
be that every portion of the genetic material of a bacterium has some 
potentiality to mature into a transmissible agent, that is, to become a 


The Interaction of Plant Viruses 
with Their Host Plants 

The information available on the interaction between host and virus 
is less extensive and less quantitative for plant viruses than for bac- 
teriophage. This is due mainly to the complexity of the hosts, to the 
difficulties involved in precise quantitative work with plant viruses, 
and also to the different interest of plant pathologists, whose main 
practical problems are diagnostic and prophylactic. We shall consider 
the growth of virus in infected plants, the changes that take place in 
infected cells, the spread of virus in the host plant, and the altera- 
tions of plant metabolism that accompany the manifestations of virus 


After tobacco leaves are infected with tobacco mosaic virus, it is 
possible to follow the amount of virus and its increase in the plant 
sap (334). If the inoculated leaves are excluded from the material 
extracted for sap, the results represent amounts of new virus formed 
in the plant. Such data, of course, describe the mass growth of virus 
in the plant as a whole and not its pattern of growth in individual cells. 
The results depend on the methodology employed. 

It had been reported that tobacco mosaic virus protein, detected by 
chemical isolation in sap from frozen plants, appeared several days 
after infection and reached the maximum after 5 to 6 weeks (623). 
With an improved method for extraction of the leaves and using elec- 
trophoretic determination of virus protein (which appears as an 
abnormal component in the electrophoretic pattern of the cytoplasmic 
protein fraction ) , virus protein could be detected 3 days after infection 
and reached a maximum after about 2 weeks (675). This early ap- 


210 Plant Viruses in the Host Plants 

pearance of virus protein in the cytoplasmic protein fraction, as distinct 
from the chloroplast fraction, is accompanied by a decrease in one of 
the major protein constituents of the same sap fraction, probably a 
nucleoprotein. Itjvas suggested (675) that the virus originates by 
direct transformation of the normal nucleoprotein into virus protein, 
without inter mediate breakdown to nonspecific building blocks. Direct 
evidence For this conclusion is lacking. It was, however, noted that 
the total amount of cytoplasmic protein remains constant, whereas the 
proportion of virus protein increases from to 403?. Moreover, virus is 
known to be formed in detached and darkened leaves, where syntBeses 
are slower (640).) The results are compatible with the idea that virus 
[synthesis is accomplished from nonspecific blocks and that the decrease 
in normal components is simply due to their normal breakdown with 
reduced replacement. In tobacco plants, as much as 10% of the dry 
weight of the plant may finally be represented by tobacco mosaic virus. 
/[Measurements of the relative radioactivity of plant nucleoprotein 
and of tobacco mosaic virus nucleoprotein (94) extracted from in- 
fected plants fed with radioactive phosphorus indicated that the virus 
nucleic acid is synthesized anew from freshly assimilated material, and 
not by transformation of plant nucleoprotein. In agreement with this 
conclusion is the observation of inhibition of tobacco mosaic virus pro- 
duction by various purine and pyrirmdine derivatives (151; 477). The 
inhibition can be relieved by normal purines or pyrimidines. Ap- 
parently, interference with utilization of purines and pyrimidines may 
block virus synthesis. This is more easily visualized in terms of syn- 
thesis of new nucleotides and nucleic acid than of transformation of 
plant nucleic acid into virus nucleic acid. 

Studies with heavy nitrogen (N 15 H 4 C1) as a tracer element in in- 
fected tobacco leaves (479), as well as analytical study on nitrogen 
metabolism (150a), indicate that virus formation proceeds directly 
from materials supplied by the medium and not from breakdown prod- 
ucts of cell proteins. They also indicate that the virus protein, once 
formed, is not further broken down and resynthesized, since no incor- 
poration of N 15 into the virus protein takes place after this protein 
stops increasing in amount. 

A microchemical technique for the isolation and determination of 
virus protein in small samples of leaf tissue ( a few square millimeters ) 
has been applied to the study of tobacco mosaic virus growth (152). 
Isolated leaves are inoculated by rubbing, and samples are cut from 
the leaves and incubated in liquid medium. Measurable amounts of 

CH. 10 

Growth of Virus 


virus are found about 72 hours after inoculation, and the maximum is 
reached after 8-10 days, when as much as 0.5 /xg of virus protein per 
square mm of leaf can be obtained (figure 70). The increase is regular 
and more or less exponential This is to be expected, whatever the 
rate of synthesis in individual cells, since what is measured is the total 
rate of virus production in a tissue infected in one or a few spots, and 
the rate of virus production reflects mainly the rate of centrifugal 
spread and invasion of new cells, a process essentially exponential in 




I i i i 





Figure 70. The amount of tobacco mosaic virus, as determined chemically, in 
samples removed from infected tobacco leaves at various times after inoculation. 
The TMV content is given in micrograms of virus protein per 302 mm 2 of leaf area. 
From: Commoner et al. (152). 

An accurate method of following virus increase in infected plants 
consists in extracting the sap from measured areas of inoculated leaves, 
clarifying it, and examining it directly in the electron microscope by 
the calibrated droplet method, by which one can count the particles of 
virus protein (33). By this method, an increase in the number of virus 
particles is observed after 18 hours, and the maximum is reached after 
4 to 5 days, with a doubling time of about 4 hours (625; figure 71). 
Exposure to 38 C suppresses virus increase rather abruptly. At the 
point of maximum growth, at least 11% of the dry weight of the plant 
is represented by virus protein. The growth curve for tomato bushy 
stunt virus is quite similar to that of tobacco mosaic virus. 

More recently, a study by infectivity measurements of the early 
phases of infection of tobacco plants with tobacco mosaic virus and 


Plant Viruses in the Host Plants 

of bean plants with tobacco necrosis virus (690b) has indicated a 
latent period of 8 to 12 hours, followed by a rapid increase in the 
amount of infectious virus, corresponding to a doubling time of about 
1 hour. The reasons for the difference between these results and 
those of previous studies (152; 625) are unknown. It seems possible 

2 3 

Days after inoculation 

Figure 71. The amount of tobacco mosaic virus, as determined by electron- 
microscope counts, in areas removed at intervals from infected tobacco leaves. 
The ordinate represents the number of virus particles per gram of leaf tissue 
(green weight). From: Steere (625). 

that the infectivity measurements may permit a study of the earliest 
parts of virus growth. The other methods may observe later, slower 
phases of virus production and miss the early phases of virus pro- 
duction because of the presence, in all samples, of an excess of 
virus particles left over from the inoculum and not participating in 

There is no evidence in any of the above studies for or against a 
sequence of virus reproduction in uninfectious form followed by 
maturation. There is some evidence, however, for a reduction in 
recoverable virus activity soon after inoculation, suggesting an eclipse 
period similar to that observed for bacteriophage. A partial disruption 
of the infecting virus particles is evidenced by the appearance of 

CH. 10 

Intracellular Manifestations 


nonviral P 32 in plants infected with P 32 -labeled tobacco mosaic virus 



In electron micrographs of infected leaf tissue, tobacco mosaic virus 
is seen in the cytoplasm, not in the nuclei (86; figure 72). It often 
appears to be intimately associated with chloroplasts. This may be 
due to a secondary invasion of these bodies. 

Figure 72. Tobacco mosaic virus demonstrated by electron microscopy of a 
thin section of an infected cell. Note the absence of virus in the nucleus. Unpub- 
lished photograph, courtesy Dr. L. M. Black, University of Illinois, Urbana. Pho- 
tographs from the same study were published by Black, Morgan and Wyckoff (86). 

Changes in the chromoprotein of the green chloroplasts have been 
correlated with virus production (191; 686). Suppression of chloro- 
plast formation and destruction of existing chloroplasts were observed 
if the infected plants were kept under conditions of nitrogen starvation, 
whereas little or no damage to the chloroplasts resulted if ample 
nitrogen supply was available. It was suggested that the chromo- 

214 Plant Viruses in the Host Plants 

protein of the plastids had a precursor in common with the virus. 
During nitrogen starvation, not only would the virus compete with the 
plastids for the precursor, but also the protein of plastids could be 
utilized for virus formation, possibly after returning to the precursor 
stage. This suggestion is hard to reconcile with the evidence for virus 
synthesis from nonspecific building blocks (479)y 

In cases of spontaneous variegations nonvirus diseases that cause 
localized reduction in chlorophyll many cells have colorless or shrunk 
chloroplasts, or have no plastids at all, resembling cells infected with 
tobacco mosaic virus ( 191; 686 ) . Cytological observations showed that 
chloroplast destruction was less and less pronounced toward the periph- 
ery of the variegated areas. The suggestion was made, therefore, that 
the variegation could be considered as somewhat "invasive," and that 
the plastids could become transformed into spreading, viruslike en- 
tities. Plastids are supposed to be related to mitochondria, and the 
possible origin of viruses from mitochondria has been suggested in 
connection with plant viruses as well as with some animal viruses. It 
is hard to see how these speculations receive any specific support from 
the experiments described above. 

Cells infected with a plant virus may or may not show pathological 
changes, depending on the virus and on the cell function (211). 
Diminution of the number and size of chloroplasts is observed in the 
"yellows" diseases. It is uncertain whether a decrease in the yellow 
pigments (carotenes, xanthophyll) is also involved. 

The most direct morphological manifestation of virus infection inside 
a host cell is the production of mtracellular inclusions, abnormal bodies 
whose presence is strictly limited to the cells of areas showing disease 
symptoms. In the tobacco mosaic group, two types of inclusions have 
Been described, the ameboid or X-bodies and the crystalline plates 
(359). The X-bodies appear early in infected cells and are formed 
by agglomeration of particles visible at first as minute granules stream- 
ing along with the cytoplasm. When fully developed, the X-bodies 
may be 5 to 40 /*, in diameter and often enclose mitochondria and oil 
globules. They have been extracted by micromanipulation; after be- 
ing washed and crushed they release infectious virus (595). If the 
cell that contains X-bodies is mechanically damaged, the bodies often 
disintegrate. It had previously been supposed that the X-bodies repre- 
sented the result of a nonspecific cell reaction to the virus, but their 
infectivity suggests that they contain virus in amounts greater than 
present in an equal volume of cytoplasm. The virus in the X-bodies 
may be combined with other cell components; tobacco mosaic virus, 

CH. 10 Intracellular Manifestations 215 

for example, has a well-established tendency to form complexes with 
a variety of proteins. 

The high virus concentration in the X-bodies is confirmed by the 
fact that in ageing cells these inclusions are often transformed into the 
other type of inclusion, the crystalline plates, whose high content in 
infectious virus is well substantiated (see frontispiece). Crystalline 
plates may be formed directly without deriving from X-bodies; they 
are variable in shape, but often show a perfect crystalline symmetry. 
They are strongly birefringent if viewed edgewise. Upon treatment 
of the cells with acids, the plates are suddenly transformed into a 
striate material, closely resembling the needles formed by crystalliza- 
tion of purified tobacco mosaic virus (54). Crystalline plates extracted 
by micromanipulation and dissolved in water yield characteristic virus 

Cytoplasmic inclusions are formed in many plant virus diseases, 
although they are not found in some well-investigated cases, for 
example, in plants infected with cucumber mosaic virus. This may be 
due to some peculiarity of the cucumber plant sap (43). Inclusions 
are never formed in the sap of the vacuole of infected cells. This sap 
probably does not contain virus. 

Intranuclear inclusions have only been observed with certainty in a , 
few virus diseases (367), among which the best known are those of 
the severe and mild etch viruses of solanaceous plants. Many nuclei 
of cells infected with etch virus contain regular plates and other crystal- 
line forms. Some of these inclusion-containing nuclei can be seen to 
divide, even in nonproliferating cells. Tobacco plants simultaneously 
infected with tobacco mosaic and mild etch viruses show, within the 
same cell, intracytoplasmic crystalline plates of the mosaic and intra- 
nuclear plates of the etch virus. This shows that two unrelated plant 
viruses can multiply in the same cell (462). 

It is clear that an understanding of virus production in the plant 
cells would be helped by studies on single intracellular cycles of 
growth, as described for bacteriophage (see chapter 8). Suitable 
techniques for similar experiments on plant viruses are not yet avail- 
able. It Would also be desirable to correlate the course of virus syn- 
thesis with the formation of noninfectious materials serologically re- 
lated to the virus (49; 473). Growth of viruses could be studied 
advantageously in plant tissues cultivated in vitro, but the techniques 
for plant tissue cultures are in certain respects not as developed as for 
animal tissues. Virus inoculation into tissues cultivated in vitro pre- 

216 Plant Viruses in the Host Plants 

sents serious difficulties. Tobacco mosaic virus has been cultivated in 
Cultures of root tips derived from virus-infected plants. 

The relation of virus infection to cell proliferation is not well under- 
stood and is probably different from case to case. On the one hand, 
certain viruses that infect cells of the reproductive tissues inhibit or 
completely suppress cell reproduction, with stunting and eventual 
death of the plant. On the other hand, virus-infected cells often 
multiply. Some of the infected cells may be stimulated to excessive 
proliferation, with the production of abnormal outgrowths called 
enations, stemming from the leaf veins. Certain virus strains are par- 
ticularly apt to cause this type of reaction, for example, the enation 
strain of tobacco mosaic virus. 

Some viruses produce true tumors in plants. The best investigated 
one (83) causes the infected plants to react to wounds by the formation 
of woody tumor tissue (figure 7). In this case the tumor response 
varies remarkably with the heredity of the plant. Another tumor virus 
is responsible for the gall-producing Fiji disease of sugar cane (503). 
The possible role of a virus in the induction of crown gall disease by 
bacteria (178a) is still a matter of speculation. 


The way in which plant viruses spread through infected host tissues 
depends on the tissues and on the portal of infection (see 43) . Viruses 
transmitted mechanically, by rubbing leaves, penetrate into a few cells 
through small wounds. Movement to neighboring cells is slow, of the 
order of a few microns per hour (654) or at most 1 or 2 mm per day 
(Ola), and depends upon virus multiplication within each cell, fol- 
lowed by transmission to neighboring cells. Transmission is believed 
to take place mainly through the intercellular connections or plasma- 
desmata. When the virus reaches the vascular tissues (phloem and 
xylem) is spreads rapidly within them (335; 336). 

The spread occurs mostly through the phloem, although spread 
through the xylem is observed, for example, with phony peach virus. 
In the phloem, the virus is carried with the food circulation, without 
necessarily multiplying in each of the sieve tube cells that it crosses. 
The rate of spreading may be as high as an inch per minute, as, for 
example, for curly top virus in the sugar beet. By way of the phloem, 
the virus may reach the roots and then the whole plant. Actively 
growing parts of a plant are generally invaded before older, mature 

CH. 10 Spread of Viruses in the Plant 217 

Fast spread starts immediately when viruses are inoculated directly 
into the vascular tissue by insect vectors. Direct infection of the roots 
with some viruses, for example, tobacco mosaic, fails to result in gen- 
eralized spread. 

I The transport of virus along with the food circulation has been 
proved by a number of experiments/ If the movement of food away 
from a shoot of an infected plant is retarded because of reduced photo- 
synthesis, either by darkness or by defoliation, the virus reaches the 
shoot faster in the ascending food current. An experiment on sugar 
beet curly top virus (60) shows this very clearly. One shoot of a beet 
plant was placed in the dark, while two other shoots were exposed to 
light. When one of the illuminated shoots was infected with curly 
top virus by means of infected insects, the shoot that was kept in the 
dark ( and in which photosynthesis and, therefore, the centripetal food 
transport were suppressed) became infected in a few days, much 
sooner than the other noninoculated shoot that was exposed to light. 
Although the virus can pass through vascular tissue without repro- 
ducing in all its cells, many cells become infected and undergo necrotic 
changes. Phloem necrosis is often widespread. It probably accounts 
for most of the circulatory disturbances in infected plants; it is some- 
times an important factor in causing their death. As a general rule, 
virus is unable to propagate across dead tissue, whether in the leaves 
or in the stem. This is shown experimentally by killing portions of the 
stem tissues through application of steam; the virus fails to go through 
the steamed regions. 

The inability of viruses to propagate through dead cells is partly 
responsible for the virus-localizing effects of the necrotic reactions 
produced in the leaves of some plants by certain virus diseases. Ap- 
parently, some cells die before transmitting the virus to neighboring 
cells, and the result is a protection of the plant against virus spread. 
Localization of virus in areas of leaf tissue may also occur without 
visible necrosis (340). In some cases a necrotic reaction is produced 
in the stem of a plant following graft with a virus-infected scion, and 
the plant dies. Certain potato varieties, when infected by grafting 
with certain strains of potato virus X, respond with a "top necrosis," 
in which the virus spreads to the stem tip and kills it. 

An interesting limitation to virus spread is evidenced by the rarity of 
virus transmission from one plant generation to the next through the 
reproductive cells. Seed transmission is rare, but has been substan- 
tiated for several viruses, among them bean mosaic and lettuce mosaic. 
Pollen transmission, if it occurs at all, is extremely rare. Prevention of 

218 Plant Viruses in the Host Plants 

see4 transmission may be due either to failure of the virus to penetrate 
the seed embryo or to the presence of virus-inhibiting substances in the 
seed; the presence of such substances has actually been verified (368). 
Seed transmission is more frequent in legumes than in other plants. It 
is possible that in these plants the modalities of oogenesis favor virus 
infection of the egg cell. The rarity of gamete transmission of viruses 
is, of course, of great significance for the survival of the plant in nature, 
since it protects the progeny of infected plants from infection. Indeed, 
virus diseases are very important economically in plants that are propa- 
gated vegetatively, by means of tubers, rhizomes, bulbs, and cuttings 
(potatoes, fruit trees, etc.). 


Out of the great wealth of information concerning the symptomatol- 
ogy and the metabolic alterations of virus-infected plants (43; 690), 
surprisingly little can be utilized for an understanding of the processes 
involved in virus reproduction or of the direct effect of virus on the 
metabolism of the host cell.if In some cases, such as those of color 
variegation (breaking) of flowers, effects on specific enzyme systems 
are probably present, but there is no evidence as to how directly the 
virus acts on the enzymes, or by what mechanism. 

It has been claimed that two main metabolic groups of virus diseases 
can be distinguished: the mosaic type, in which the nitrogen content 
of the plant is increased and the carbohydrate content diminished; and 
the yellows type, in which the opposite changes are observed. But 
many exceptions are reported. In several cases there is an increase 
in the starch content of the leaves, but it is difficult to say whether this 
is due to increased production, reduced utilization, or reduced transfer 
of starch. Reduction of photosynthesis is accompanied, in potato leaf 
roll, by an accumulation of starch and a reduced transportation of sugar 
as compared with the normal plant (40). It is not known whether the 
initial lesion responsible for this is an alteration of photosynthesis, or 
a lesion of the phloem, or possibly an alteration of cell enzymes in- 
directly concerned with starch metabolism. The accumulation of 
starch often results in increased cell respiration. The changes in nitro- 
gen metabolism in tissues infected with tobacco mosaic virus seem to 
be due to the actual process of virus synthesis, rather than to an altera- 
tion of the nitrogen metabolism of the host (150a). 

The effects of the nitrogen and phosphorus content of soil on tobacco 
mosaic virus, synthesis seem to be exerted mainly through effects on 

CH. 10 Viruses and Plant Metabolism 219 

plant growth (45). Attempts to curb plant virus infections with anti- 
biotics effective against bacterial diseases of man and animals have 
yielded, as expected, negative results (55). 

Increased photosynthesis immediately before mechanical inoculation 
of leaves with various viruses reduces the incidence of infection (51). 
It is not known whether this effect is exerted through the host me- 
tabolism or simply through changes in turgor and vulnerability of the 
cells, which have to be wounded for the virus to penetrate. 


The Interaction of Animal Viruses 

with Their Hosts Tissue Cultures 

Intracellular Inclusions 


The study of the interaction of animal viruses with their hosts is 
complicated by the high degree of host differentiation. Structurally, 
an animal consists of highly specialized organs and tissues. The ex- 
ternal layers consist of modified cells impenetrable to most parasites. 
The spread of a* virus in the tissues is dependent on complex systems 
of circulation and also, sometimes, on propagation along differentiated 
channels such as nerve fibers. Serological mechanisms of immunity 
add to the complexity of the virus-host interaction in animals. 

As a rule, manifestations of virus infection at the cellular level can 
be traced more directly in systems such as tissue cultures, which con- 
tain cells of more or less similar genetic and developmental make-up, 
or in simple, accessible tissues such as the corneal epithelium. More 
complicated systems, such as the fertilized chicken egg, also provide 
highly reproducible systems for the study of virus reproduction in single 
layers of living cells. 

It should be the goal of a sound biological approach to interpret the 
complex situations that arise in virus infections of man and animals in 
terms of the elementary virus -cell interactions analyzed in simpler 
situations. This goal can seldom be reached in a satisfactory manner. 
Yet, the careful analysis of the course of infection in the adult animal 
body and of the virus spread in animal populations may throw some 
light on basic virus properties (see 215). For example, the rather 
precise incubation time or death time of certain virus infections, and 
the dependence of these times on the amount of virus inoculated in 


CH. 11 

Growth of Viruses in Animals 


experimental infections, suggest that the response that is observed 
requires the presence of a fairly precise concentration of virus (248). 
From survival data, therefore, it may be possible to estimate the rate 
of virus multiplication in the animal independently of any information 
as to virus localization. 


1 1 

r . x T 

Experiment II 
Adapted Rhodes virus _ 
o Unadapted Rhodes vrrus 

1 i 1 


10 1 - 

6 12 24 48 72 

Hours after intranasal inoculation 

Figure 73. The amount of influenza virus A in ground lungs of mice after intra- 
nasal inoculation. The greater reproduction of a mouse-adapted strain is com- 
pared with the less extensive multiplication of a strain passaged in ferrets and 
eggs. The infectious titer is given in EID r , 's ( 1 EID f)0 = dose infecting 50% of 
inoculated eggs). From: Davenport and Francis (163). 

Growth of viruses in the animal host. Many authors have reported 
growth curves of viruses in the intact animal, for example, for influenza 
viruses in the mouse lung (163). Groups of animals are sacrificed at 
intervals, and the lungs are pooled, ground, and titrated. The growth 
curves are of the type shown in figure 73. The generation times can 
be calculated; figure 73 gives a generation time of the order of 1 hour. 
These results and others of the same type, however, are probably a 
rather distant picture of the process of virus reproduction at the cell 
level. Rather, they reflect complex cycles of virus infection, repro- 
duction, liberation, reinfection, combination with inhibitors, and so on. 

An example of a more direct approach is provided by experiments 


Animal Viruses Tissue Cultures 

on pneumonia virus of mice in the mouse lung. These experiments 
were directly aimed at testing the existence of a growth cycle similar 
to that of phage in sensitive bacteria and the relation of this cycle to 
the development of pulmonary lesions (254; 255; 347). Recoverable 
virus infectivity decreased to almost zero a few minutes after inocu- 

1024 *-HI i i i i i i 

x Hemagglutination 


- -3.0 



12 18 24 30 
Hours after PVM 

Figure 74. The amount of pneumonia virus of mice (PVM) in mice lungs 
ground at intervals after intranasal inoculation of virus. From: Ginsberg and 
Horsfall (255). 

lation (eclipse period) and remained low for 12-18 hours. Then, a 
rise by a factor of 10-20 occurred in 10-15 hours, followed by a plateau 
and by further rises (figure 74). Thus, one cycle of latent period and 
liberation would take about 30 hours. Of course, the apparent yield 
in such a cycle is bound to be a minimum estimate, since readsorption 
of the new virus, with new eclipses, must occur immediately upon 
liberation. The development of pneumonic lesions closely parallels the 
rise in virus titer. The degree of suppression of the pneumonic symp- 
toms by the capsular polysaccharide of Klebsiella pneumoniae, a 
specific inhibitor for pneumonia virus of mice, is -also in direct pro- 
portion to the suppression of viral multiplication. When, in the 

CH. 11 Factors Influencing Animal Infections 223 

second week of infection, the virus titer begins to decrease in the 
lung, no further development of lung lesions takes place. 

At least one feature of the pattern of virus production just described, 
the eclipse period, seems to be of general occurrence in virus infections 
(550). Nonrecoverability of virus soon after inoculation has been re- 
ported for rabies virus, poliomyelitis, St. Louis encephalitis, yellow 
fever, influenza, and several others. 

Factors influencing animal virus infections. A set of empirical ob- 
servations of great potential significance concerns the influence of the 
diet of the host on the course of viral infections, particularly with 
neurotropic viruses (134). A variety of vitamin, amino acid, and min- 
eral deficiencies have been found to result in milder infections, with 
delayed or reduced lethality or with changes in symptomatology. The 
mechanisms of these effects are unknown. One may speculate on 
specific requirements for the deficient substances by different synthetic 
systems, so that in deficient animals virus synthesis may be interfered 
with, in the absence of irreparable damage to the tissues. An interest- 
ing observation is that eggs from riboflavine-deficient hens support 
growth of a virus (blue tongue of sheep) which fails to grow in normal 
eggs (476). 

(Suppression of virus growth in animals by chemicals has been re- 
ported repeatedly.) Many of the reports concern mainly the thera- 
peutic or preventive effect of chemicals on virus diseases. /Among the 
few examples of successful virus chemotherapy, we have the action of 
sulfonamides, penicillin, chloromycetin, and aureomycin on viruses of 
the psittacosis-lymphogranuloma group (162). A naphthoquinine de- 
rivative protects mice against the neurotropic virus Col SK (586). Cer- 
tain pyrimidine analogues can suppress vaccinia virus in mice (648a). 
Helenine, a fungal product, protects mice against some viruses (602a). 
In all these cases, the mechanism of action is unknown, and we have 
no information as to what phase of virus production is affected, jf 

Toxic manifestations, which are observed too early after virus inocu- 
lation to be attributed to virus proliferation, have been described for 
viruses of the lymphogranuloma (544) and of the influenza groups 
(162a; 298). The lymphogranuloma-type toxins may be of the nature 
of endotoxins, since they are not readily separable from the virus 
particles. For influenza, there is evidence that the toxic reaction is 
due to an abortive infection of cells in which virus fails to reproduce 
to maturity, for example, in the cells of adult mouse brain ( 582 ) . Toxic 
symptoms are also produced by intravenous or intraperitoneal inocula- 

224 Animal Viruses Tissue Cultures 

tion of influenza viruses and by inoculation of Newcastle virus in mouse 
lung, where this virus does not multiply, at least in infectious form 


Tissue-culture techniques. The principle of tissue cultures is based 
on the classic observations made in 1907 by Harrison. If fragments 
of tissues from a living host are isolated with sterile precautions and 
placed in a nutrient solution, the cells not only continue to metabolize 
but often divide, particularly if the cultivated tissue fragment came 
from an embryo (generally, 7-10-day-old chick embryos are used). 
The reduced geometrical and chemical interaction among tissues in 
cultures may actually produce an increase in the rate of cell growth, 
a return to reproducing habits in cells that had stopped multiplying, 
and a loss of some of the characteristics of the cells, which tend to 
manifest more embryonal characters than they did when still in the 

All the microscopical culture methods, introduced mainly by Har- 
rison, Carrel, and Fisher (see 287; 577; 655), involve the cultivation 
of small fragments of tissue in balanced salt solutions ( Tyrode * or 
others) to which sugar, blood plasma and extracts of fresh chick em- 
bryos are added. Serum or serum ultrafiltrate from the same animal 
species that provides the tissue is generally added. The plasma co- 
agulates and immobilizes the fragments on a cover slip. Proliferation 
of cells takes place in the thin layer of coagulated plasma and the 
growing cells migrate along the surface of the cover slip, so that direct 
microscopic observation is possible. The cover slip can be coated with 
a collodion layer. This can later be lifted off with a thin layer of cells 
and mounted for observation in the electron microscope. 

Harrison's original method, using hanging-drop cultures in a depres- 
sion slide (figure 75), requires the transfer of the fragments of tissue 
to a new culture every 2 or 3 days to insure continued development. 
The whole procedure must be carried out under rigorous sterile pre- 
cautions, since bacterial contamination would quickly cause complete 
disintegration of the culture. Antibiotics may be added to the cultures 
to suppress bacterial contaminants. Carrel's method avoids some of 
the difficulties of the slide technique by the use of a special flask ( fig- 
ure 75), one wall of which can be made of cover slip thickness, in 

iTyrode's solution: NaCl, 8 grains; CaCl 2 , 0.2 gram; MgCl 2 , 0.2 gram; KCl, 
0.2 gram; Na 2 CO 3 , 1 gram; NaH 2 PO 4 , 0.05 gram; water, 1000 ml. Adjust to pH 
7.4 using H 3 PO 4 and Na 2 CO 3 . Sterilize by filtration. 

CH. 11 

Viruses in Tissue Culture 


which relatively larger amounts of tissue can be cultivated for a long 
time with frequent changes of nutrient medium. 

Virus infection in cultures of this type can be produced either by 
infecting the tissue prior to setting up the culture or by addition of 
virus to the culture fluid. A useful modification of the flask technique 

Figure 75. Containers used for tissue cultures. 1. Grand's flask. 2. Porter's 
roller flask. 3. Roller tube with flattened surfaces. 4, 5, Carrel's flasks. 6, 7. 
Depression slides, on which cultures on coverslips are inverted. 8. Earle's flask. 
From: Cameron, Tissue Culture Technique, 2nd Ed., Academic Press, New York. 
Courtesy, Dr. G. Cameron, Medical Research Foundation of Dade County, Miami, 

is the "roller-tube" method (282), in which the tissue fragments are 
immobilized by plasma on the walls of test tubes. These receive a 
nutrient solution and are slowly rotated in a drum, so that oxygenation 
of the fluid can be kept at a high level. Thus, longer intervals of 
growth without transfer are made possible. An agar-slant technique 
for tissue culture has also been used for virus cultivation (508). 
Chemically defined media for the cultivation of tissues have been 
developed (508b). 

For the specific purpose of virus culture, simplified techniques have 
been developed, which involve the cultivation of tissue fragments 
floating freely in a thin layer of nutrient fluid (423; 469). This type 

226 Animal Viruses-Tissue Cultures 

of culture can best be used for quantitative studies of virus growth, 
but does not allow for microscopical observation of the cells in the 
living state. It is really a tissue-survival rather than a tissue-cultivation 
method, since hardly any cell reproduction takes place, at least for most 
tissues, and the viruses reproduce in the metabolizing but nondividing 

The tissues employed in cultures to be used for virus cultivation vary 
according to the virus and to the purpose of the experiment; practi- 
cally all kinds of tissues have been used. It should be recalled that 
certain tissues, such as differentiated nerve tissue, muscular tissue, and 
many glandular epithelia, fail to reproduce in tissue culture of any 
type, although their cells may remain alive and metabolize actively for 
long periods. 

Pure cultures of certain cell types can be set up either by isolating 
layers of similar cells before placing them in the culture vessel (mainly 
feasible for surface epithelia ) or by utilizing the differential prolifera- 
tion rate of different types of cells in slide or roller-tube cultures 
(124). For example, if a slide culture of connective tissue is allowed 
to grow for a few days, the fibroblasts may move into the peripheral 
region faster than, for example, the macrophages. After several trans- 
fers, it often becomes possible to isolate a group of cells of one type 
and to transfer them to a new culture, which will then contain cells 
of one type only. 

A recently developed method (196; see figure 4) allows the study of 
local lesions or plaques formed by some viruses on a layer that consists 
essentially of a homogeneous type of cell. Many fine fragments of 
strained embryonal tissue are grown on a layer of culture fluid. The 
fibroblasts, which are the fastest-growing cells, rapidly form a con- 
tinuous layer, The fluid is removed, the cell layer is washed, and virus 
is applied. Then a layer of nutrient agar is added. This immobilizes 
the fibroblast layer and allows the virus to form localized lesions, easily 
recognizable by naked eye or with low magnification. 

Growth of viruses in tissue culture. The possibility of obtaining 
virus reproduction in tissue cultures was quickly recognized. The 
technique became quite common for the preparation of relatively large 
amounts of viruses in a relatively simple medium. For example, a 
small inoculum of vaccinia virus in a flask culture will yield up to 
10 5 times as much virus within 48 hours. This virus is free from 
bacterial contamination and is very convenient for the preparation of 

CH. 11 Viruses in Tissue Culture 227 

The need for actual contact of the virus with the tissue as a pre- 
requisite for virus growth was shown clearly by experiments in which 
the virus was separated from the tissue by collodion membranes that 
prevented the passage of virus particles; no virus reproduction took 
place (495). Killed cells fail to support virus growth. In general, the 
virus grows in cultures in which conditions are favorable for the pro- 
liferation of host cells. This can be tested in slide cultures, by com- 
paring the extent of virus production either with the number of mitotic 
figures present in the culture or with the extent of tissue growth, as 
judged by the size of the zone of cellular migration. Virus growth is 
not always tied up with cell proliferation, however. For example, 
fowl-plague virus grows well at a temperature of 30 C in chick em- 
bryo tissue, which at this temperature shows very few mitotic figures 

Host and tissue specificities can be studied in cultures of different 
tissues from different animals. The specificities are not always iden- 
tical with those manifest in the animal body, but the causes of dif- 
ference are often trivial. For example, virus III of rabbits grows only 
in cultures of rabbit testicle (24), although in the animal it grows in 
other tissues, whose cells, however, survive and proliferate less well in 
tissue cultures. Several viruses can grow in cultures of tissues from 
animals that do not generally support their growth. Thus, vaccinia, 
yellow fever, influenza, and many other viruses reproduce in tissues 
from the chick embryo, although they fail to grow in adult chickens. 
In some instances the difference is easily explained; influenza, for 
example, is unable to grow at the high temperature of the bird's body 
(207), but grows both in embryonated eggs at 37-38 C and in cul- 
tures of chick embryo tissues. 

Tissue specificity, as well as host specificity, is often preserved in 
tissue cultures. Thus, foot-and-mouth virus reproduces only in epi- 
thelial cultures from horse and cattle. Rous sarcoma virus grows in 
cultures of macrophages better than in fibroblasts (124). In macro- 
phage cultures it can induce a true cancerization. Poliomyelitis virus 
can be grown in cultures of extraneural tissues, especially fibroblasts, a 
fact that agrees with evidence for the multiplication of this virus in 
extraneural tissues in the infected animal host (577 a; 672). Different 
strains of host cells cultivated in vitro may show a variety of different 
types of response to the same virus (37a). 

The new host-range potentialities revealed by tissue cultures are 
probably due to the use of embryonal tissue. The changed specificity 
relations between host and virus in embryonal tissues is one of the 

228 Animal Viruses Tissue Cultures 

most interesting problems of virology, but one that still needs clari- 
fication. It would seem likely that one of two factors is involved: 
(a) the lack in the embryo of inhibitory substances that appear late 
in the course of development; (b) the presence in embryonal tissues 
of certain enzyme systems or substrates needed for virus growth and 
lacking or inoperative in adult tissues. 

Next to the problem of specificity, we must consider what tissue 
cultures tell us concerning the mode of virus reproduction. Informa- 
tion of two kinds is available, the first derived from virus titration, the 
other from direct microscopic observation of abnormal manifestations 
in infected cells. Unfortunately, little quantitative evidence on virus 
multiplication can be obtained by either of these methods, particularly 
because the relation between the amount of tissue present in a culture 
and the extent of virus multiplication is complicated by innumerable 
factors. From separate titrations of virus in the fluid and in the tissue 
fragments in flask cultures infected with psittacosis (455), foot-and- 
mouth, and other viruses, we have learned that the virus introduced in 
the culture disappears from the liquid phase and is taken up by the 
tissue. After 1 or 2 days, during which the amount of recoverable virus 
activity increases in the tissue without being liberated, virus reappears 
in the fluid phase and is later taken up again by the tissue. The ob- 
vious similarity of these stages with the cycles of growth of bacterio- 
phage in a bacterial culture suggests similar cycles of intracellular 
growth and liberation. After virus production is completed, virus may 
remain active in a culture for various lengths of time, up to several 
months (214). 

Tissue-culture experiments have provided evidence for the actual 
penetration of a virus inside host cells. Rons and his coworkers (567) 
prepared suspensions of macrophages taken from slide tissue cultures 
by tryptic digestion of the coagulated plasma. Vaccinia or fibroma 
virus, mixed with the cell suspensions, was quickly taken up by the 
cells and apparently penetrated inside them. If the infected cells were 
mixed with antivirus serum and the mixture was injected into a sus- 
ceptible animal, virus infection followed, which proved that the virus 
had been protected against the action of the antiserum. These experi- 
ments, using suspensions of animal cells prepared from tissue cultures, 
illustrate the potential applications of such a technique, which, unfortu- 
nately, has long been neglected in the analysis of host-virus interaction. 
A new application of this method to the cultivation and diagnosis of 
poliomyelitis virus, using cells from cultures of a human cancer, prom- 
ises very interesting results (577b). 

CH. 11 Viruses in Tissue Culture 229 

Viruses have been found to grow in cultures of tissues from animals 
immunized against them (24). This shows that acquired immunity is 
essentially due to antibodies and not to any intrinsic changes in the 
cells of the host. Protection of tissue cultures by antivirus serum given 
before the virus or together with it seems to be of rather general occur- 
rence (288), although it is not always as complete or as permanent as 
in the inoculation of virus-antiserum mixtures into an animal, where 
the virus is probably destroyed rapidly if prevented from growing 
(226). Viruses that are susceptible to drugs-for example, lympho- 
granuloma to sulfonamides-may be suppressed by drug treatment in 
tissue cultures as well as in the animal body (283). 

Tissue cultures have been used for some studies on the relation of 
virus growth to the level of host metabolism. Zinsser and Schoenbach 
(697) found that in cultures of chick-embryo brain the increase in 
equine encephalomyelitis virus occurred mainly in the first days of 
culture, when the oxygen uptake was highest. Later, oxygen consump- 
tion and virus reproduction diminished in parallel. In contrast, typhus 
rickettsiae multiplied most actively after the oxygen uptake had ceased 
and when presumably the cells were no longer viable. This and other 
observations (468) indicate that animal viruses, just as bacteriophage, 
utilize for their reproduction the enzymatic machinery of the host cell, 
rather than simply using preformed substrates. 

Observations on the effect of various chemicals on virus growth in 
tissue cultures are in agreement with this general viewpoint. Meta- 
bolic poisons (such as cyanide, dinitrophenol, malonate) interfere with 
the formation of new virus, probably by preventing the flow of energy 
through the proper cellular channels (1 ). A variety of other chemicals 
also suppress or reduce virus growth. In screening for compounds of 
potential chemotherapeutic value in virus infections, tissue cultures 
are widely employed. Even compounds that only reduce or retard 
virus growth in tissue cultures are worth considering as possible 
chemotherapeutic agents, since in the intact animal body a reduction 
of virus growth will increase the chances for successful body defense 
by means of antibodies and other defense mechanisms. Substances 
that reduce virus production when administered after the virus are 
considered especially worthy of study. 

From the standpoint of interpreting the effect of chemicals in terms 
of the mechanism of virus growth, few studies on tissue culture have 
yielded information of interest. Virus production has generally been 
determined only by the final virus titer, rather than by series of meas- 
urements of virus in the tissue and in the surrounding fluid. The most 

230 Animal Viruses Tissue Cultures 

interesting results concern the effect of analogues of amino acids and 
of nucleic acid constituents. For example, DL-ethionine, an analogue 
of methionine, inhibits the growth of influenza virus in cultures of 
chorioallantoic membrane of chick embryo (2). Purine analogues, 
such as 2,6-diaminopurine, suppress production of vaccinia virus in 
minced-tissue cultures (648). The effect of these chemicals can be 
counteracted by the corresponding metabolites ( for example, methio- 
nine for ethionine; adenine for 2,6-diaminopurine). How direct the 
action of these inhibitors on virus production is, we cannot say. It is, 
however, a reasonable hypothesis that they interfere with virus pro- 
duction by blocking the utilization of the corresponding normal me- 
tabolites for the synthesis of virus protein or virus nucleic acid. This 
would then indicate that production of animal viruses involves the 
synthesis of new protein and nucleic acid from simple building blocks 
rather than the transformation of large molecules of cell protoplasm 
into virus material. The reasons for the suppressive effect of certain 
basic amino acids on some viruses (500) remain obscure. 


Microscopic study of viruses in tissue cultures cannot always provide 
evidence as to the extent of infection. Some viruses, like equine en- 
cephalomyelitis, cause extensive death of cells; others may multiply 
abundantly without visible cell destruction, probably because of the 
ability of infected cells to divide repeatedly, transmitting virus to their 
daughter cells. 

Microscopic examination often allows us to recognize, inside the 
infected cells, the elementary bodies of those viruses whose particles 
can be. visualized either by staining methods or in the dark field ( see 
chapter 4). The major worth of microscopic observation of virus- 
infected animal tissues resides in the study of the so-called virus in- 
clusions (217; 220). 

It was noticed early by pathologists that the cells of tissues infected 
with several diseases, now known to be caused by viruses, showed 
rather characteristic bodies (inclusion bodies) not visible in uninfected 
cells. As early as 1869 such bodies were observed by Rivolta in tissues 
of chickens suffering from fowl pox. In 1892 Guarnieri discovered 
typical inclusions in cells infected with vaccinia or smallpox virus, and 
in 1903 the Negri bodies were discovered in the brains of rabid dogs. 

These inclusions are much larger than the elementary bodies of the 
corresponding viruses. They are easily recognizable microscopically 

CH. 11 Intracellular Virus Inclusions 231 

(figures 76-78) and can grow to the extent of occupying most of the 
host cells and of pushing cell structures aside (see figure 76). Such 
inclusions can be found almost without exception in tissues infected 
with a given virus, no matter in what tissue the virus is growing. They 
are formed either in the cytoplasm or in the nucleus of infected cells, 
and their location is generally specific for a given virus. 

At first, the inclusions were mistaken for protozoa. This mistake 
gave rise to the term Chlamydozoa (mantle animals) to designate these 
agents of disease (Von Prowazek). As the viral nature of the agents 
of the diseases became firmly established, the relation of the intra- 
cellular inclusions to the virus elements that carry the infectivity in 
the extracellular state became a matter of debate. Some authors con- 
sidered the inclusions as products of the reaction of the infected cell 
to the virus; others considered them as actual virus colonies. In spite 
of the support apparently given to the first hypothesis by the occasional 
presence in uninfected tissues of inclusions similar to those associated 
with virus diseases, most workers today agree that most or all inclusions 
result directly from the reproduction of virus and contain elements 
which after isolation can be shown to carry virus activity. It has not 
yet been possible, however, to decide how the inclusions are formed in 
each individual case. 

Some inclusions are compact enough to permit extraction from the 
cells by micromanipulation, followed by washing, crushing, and testing 
for virus activity. In this manner, the inclusions of molluscum con- 
tagiosum (which are surrounded by a tough membrane, probably a 
reaction product of the infected cells ) and those of fowl pox ( Bellinger 
bodies ) were shown to contain several thousand infectious elementary 
bodies (262). The bodies appear to be embedded in a matrix, prob- 
ably containing ribonucleic acid. 

Some cytoplasmic inclusions are of great diagnostic importance. For 
example, the diagnosis of rabies in dogs is made by the finding of 
intracellular inclusions ( Negri bodies ) in the cells of Ammon's horn of 
the brain. The Negri bodies are also found in epithelial cells of the 
infected rabbit cornea. Cells containing Negri bodies can go through 
the mitotic process. 

Development of virus inclusions. If the inclusions contain the virus 
and result from its growth, it becomes particularly interesting to follow 
the process of their formation. Interesting observations have been 
made by tissue-culture methods, supplemented by observations on 
tissues of infected animals. 

232 Animal Viruses Tissue Cultures 

The virus of psittacosis has been thoroughly investigated in tissue 
cultures of fibroblasts and of lung epithelium (89; 418). A first phase 
without identifiable virus is followed after 12 to 24 hours by the ap- 
pearance of "homogeneous plaques," larger than elementary bodies. 
These plaques increase in size and assume a more granular appearance, 
leading to the formation of large colonies of elementary bodies; the 
cells are then ruptured and the elementary bodies are dispersed out- 

Figure 76. Section of a molluscum contagiosum lesion from human skin. Note 
the almost complete transformation of the cells into inclusion bodies and amorphous 
debris. Courtesy Dr. H, Blank, University of Pennsylvania, Philadelphia. 

side. As a counterpart to these appearances, we may notice that the 
agents of mouse pneumonitis and feline pneumonitis, two viruses be- 
longing to the psittacosis group, have been reported to possess a stage 
of extracellular reproduction (671). This claim is open to serious 

Elementary bodies of vaccinia were already seen by Buist, Paschen, 
and other workers inside infected cells before their identity with the 
virus agent was proved by studies on isolated and purified bodies. 
Borrel (96) made his classic studies of the elementary bodies of mam- 
malian and fowl-pox viruses in the fluid from pox blisters, which con- 
tain many cells loaded with elementary bodies. Several workers 
analyzed the appearance and numbers of these bodies from infected 
tissues, using all sorts of staining methods. More fruitful were studies 
in which the fate of elementary bodies and the development of in- 

CH. 11 Intracellular Virus Inclusions 233 

elusions were studied by direct observations in tissue cultures. Bland 
and Robinow ( 90 ) described in detail the development of inclusions at 
various intervals after accurately timed infection with washed vaccinia 
particles. At first, the particles were seen unchanged in the cells; a 
few hours later, the elementary bodies were no longer visible, and in 
their place larger bodies appeared. Still later, these gave place to less 
compact and less stainable bodies, the "networks," which, like the 
elementary bodies, gave a positive Feulgen reaction. The networks 

Figure // (lett;. intraceiuuar inclusions or psittacosis \irus in tissue cultures 
of chick-embryo fibroblasts. Dark-field microscope. The arrows indicate two 
virus inclusions containing elementary bodies. From: Bland and Canti (89). 

Figure 78 (right). Vaccinia virus inclusion in a corneal cell of rabbit. Dark- 
field microscope. Note the elementary bodies. From: Eisenberg-Merling, J. 
Path. Bact., 50:279, 1940. 

increased in size and after 1 or 2 days often split up into a number 
of small granules, which were scattered around in the cytoplasm, thus 
apparently returning to the stage of elementary bodies. A direct 
morphological continuity between the infecting elementary bodies and 
the networks is not proved, and it is possible that a phase without 
microscopically recognizable virus elements intervenes. 

A number of other descriptions of the microscopic events in cells 
infected with vaccinia virus concur in showing that the inclusions are 
related to the infecting elementary bodies by a series of complex mor- 
phological transformations, which suggest a life cycle. Ultimately, this 
gives rise to a great many elementary bodies. In some cases, however, 
the elementary bodies were reported to be present as such throughout 
the process (321; 322; 643). They simply increased in number and 
underwent slight morphological changes, ultimately filling the cell com- 

234 Animal Viruses Tissue Cultures 

pletely and causing it to burst and to release the virus particles. The 
different results may in part have been due to differences in tissue, 
since different observations were made on the corneal and the chorio- 
allantoic epithelia. The differences might depend on the mode of in- 
fection, on the amount of virus inoculated, or on some other unknown 
factor. The healthy or damaged conditions of individual cells may also 
play a role. 

The electron microscope provides some evidence on the formation 
of virus inclusions. Virus-infected cells from tissue cultures were used 
first for electron microscopy; the thin layers of cytoplasm of the 
periphery of cells that spread over the cover slip in slide cultures (528) 
were used. These thinly spread cells can be fixed, lifted on a collodion 
membrane, and observed in the electron microscope. Particles of some 
viruses could easily be observed because of their high electron scatter- 
ing power, as shown in figure 79. Virus particles in these thin layers 
of cytoplasm were often found in pairs. This may simply reflect the 
tendency of spherical particles to run together when drying in a thin 
layer of semifluid medium. Destructive viruses such as equine en- 
cephalomyelitis may transform cells into practically pure masses of 
virus particles (figure 80). 

The introduction of perfected sectioning techniques has made it 
possible to cut tissues into slices thin enough for direct electron 
microscopy. Several virus inclusions have been studied by this method, 
but the observations are not yet fully interpreted. Chorioallantoic 
membranes of chick embryo infected with fowl pox show all stages 
between clusters of elementary bodies and large inclusions in which 
elementary bodies are not recognizable as such (490). In tissue cul- 
tures infected with fowl plague virus (225) the electron microscope 
shows nuclear damage, especially in the nucleoli, accompanied by for- 
mation of extranuclear filaments, which seem to split into elementary 

Examination of vaccinia virus in the chick-embryo membrane shows 
the formation of elementary bodies, in a more or less homogeneous 
matrix, with transition forms between the homogeneous stage and the 
differentiated, fully formed elementary bodies. The latter are observed 
in practically pure form in the most advanced stages of cell infection 

Rather similar observations have been reported on molluscum con- 
tagiosum (36). Here, the typical inclusions seem to evolve by the 
formation of a nucleic acid-containing matrix, within which new virus 
particles are formed by a stagewise process from large, barely recog- 
nizable condensations of material in the homogeneous matrix to the 

CH. 11 

Intracellular Virus Inclusions 


fully developed virus particles (figure 81). Some of the particles, 
when sectioned by the microtome knife, appear to be hollow. 

These observations on the formation of intracellular inclusions, taken 

Figure 79. A portion of a cell of chicken tumor I ( Rous sarcoma) cultivated 
jn vitro. Note the virus particles about 70-85 m/u, in diameter in the thin layer of 
protoplasm. From: Claude, Porter, and Pickels, Cancer Res. 7:421, 1947. Cour- 
tesy Dr. K. R. Porter, Rockefeller Institute, New York. 

as a whole, support the idea that reproduction of virus material takes 
place by a process in which mature virus particles are not recognizable. 
The mature particles appear to be formed as a terminal stage in virus 
production. Thus there seems to be a further analogy with bacterio- 
phage reproduction, besides the occurrence of the eclipse of infectious 


Animal Viruses Tissue Cultures 

Figure 80. Eastern equine encephalomyelitis virus in a disintegrating cell of 
chick-embryo tissue culture. From: Bang, Ann. N. Y. Acad. Set., 54:892, 1952. 
Courtesy Dr. F. B. Bang, Johns Hopkins University, Baltimore. 

(a) (b) 

Figure 81. Formation of particles of molluscum contagiosum virus in an inclu- 
sion body, (a) Virus particles in locules surrounded by a cytoplasmic matrix, in 
which the particles appear to be formed, (b) A detail from a similar section, 
showing what may be immature virus particles in the matrix. | From: Melnick 
et al., Ann. N. Y. Acad. Sci., 54:1214, 1952. Courtesy Dr. J. L. Melnick, Yale 
University, New Haven. 

CH. 11 Intracellular Virus Inclusions 237 

virus following infection. Eclipse of infectivity and late formation of 
mature virus particles would be two aspects of the series of trans- 
formations of the viruses in their intracellular life cycle. 

Intranuclear inclusions. In a number of virus infections, such as 
yellow fever, herpes, chicken pox and infectious warts, inclusions are 
found inside the cell nucleus rather than in the cytoplasm. Intra- 
nuclear inclusions are present in tissue cultures as well as in the tissues 
of diseased animals. These inclusions, which are more acidophilic than 
the nuclear material, have been divided into type A and type B. Type 
A is accompanied by a greater disruption of nuclear chromatin, type B 
by less destruction. No observation of mitosis in cells with intra- 
nuclear inclusions seems to have been reported. 

The relation between intranuclear inclusions and virus particles has 
not been investigated thoroughly. The inclusions in the nuclei of cells 
infected with herpes virus contain basophilic granules, which may 
represent the virus particles themselves. Virus has been demonstrated 
in herpes inclusions isolated by micromanipulation (42). Yet the 
nuclei do not seem to contain more virus than the rest of the cell 
(232). Relatively large amounts of herpes virus are found, instead, 
in the mitochondria! fractions of the cells (4). Virus might be pro- 
duced first in the nucleus and then migrate into the cytoplasm. It has 
been reported (159) that the Feulgen-positive inclusions of herpes in 
the chorioallantoic cells of chick embryo become progressively less 
stainable as the infection proceeds, and the nuclei become poorer in 

Some virus-induced papillomas of man contain characteristic intra- 
nuclear inclusions and yield virus particles with a tendency to form 
regular, crystallike patterns (633). The presence of these particles 
in the intranuclear inclusions has not yet been established. 

A particular group of intranuclear inclusions are the "polyhedra" 
found in a number of infectious diseases of insect larvae. These crystal- 
like inclusions contain both the virus, in the form of rod-shaped par- 
ticles and their presumptive precursors (68a), and a noninfectious 
protein serologically related to the virus (see chapter 5). The nu- 
clear chromatin is destroyed, and the nuclei become completely full of 

In tissue cultures of ovarian tissue of silkworm infected with silk- 
worm jaundice (649), polyhedra appear 3 hours after inoculation in 
the cells that line the ovarian tubules; these are the cells that proliferate 
best in the cultures. It is interesting that the adult moth, whose tissues 
are derived in part from just these cells, is quite resistant to the virus. 

238 Animal Viruses Tissue Cultures 

Multiple inclusions. Inclusions have been found to occur in the 
nucleus and in the cytoplasm of the same cell in mixed infection with 
two unrelated viruses, one of which gives intranuclear, the other cyto- 
plasmic, inclusions (14; 637). Evidently, a cell already infected with 
one virus may become infected with a second one. Either a nuclear 
dysfunction is not necessarily present in intranuclear infection, or, if 
present, it is not incompatible with the development of another virus 
in the cytoplasm of the same cell. This in turn may suggest that 
growth of a virus may not require the functions of the nuclear material 
of the host cell, but only the functions of the enzymatic machinery, in 
the same way as was suggested for phage growth (see chapter 8). We 
must remember, however, that most cells of animal or plant tissues 
are more complex in organization than bacterial cells, and that the 
machinery needed by a virus or disrupted by it may only be a small 
part of the available cell machinery. It would be quite important to 
clarify the conditions that make some viruses grow in the cytoplasm 
and others in the nucleus of the host cells; this might also throw light 
on specific biochemical functions of these cell constituents. 

Swedish workers have analyzed by cytochemical procedures (see 
127) the amount and localization of protein and nucleic acids in virus- 
infected cells (357). Viruses that contain DNA cause an increase of 
protein synthesis and an accumulation of DNA, either in the cytoplasm 
(molluscum) or in the nucleus (infectious warts), wherever the virus 
produces inclusions. Neurotropic viruses which supposedly contain 
RNA (poliomyelitis, louping ill) cause the appearance of intranuclear 
granules containing protein and RNA (see also 540). The hypothesis 
was put forward that viruses which contain RNA are genetically less 
complex and consist of a single type of nucleoprotein, resembling 
cytoplasmic nucleoproteins, whereas the DNA viruses are genetically 
more complex and more similar to chromosomal material. 

This hypothesis seems to go beyond the evidence. In the first place, 
the localization of virus particles or inclusions in the same cells on 
which the cytochemical analyses were carried out was inadequate. In 
the second place, there are no grounds for the supposed difference in 
genetic complexity of DNA- and RNA-containing viruses. Finally, 
the presence of RNA in some of the viruses used in the experimental 
work is doubtful. 

An increase in the amount of nucleic acid in tissues of chick embryo 
infected with herpes virus has been reported (3), together with specific 
changes in certain enzymatic activities. We do not know how directly 
these changes are brought about by the virus. 



Growth of Viruses 
in the Chick Embryo 

A very convenient host for the experimental study of virus infection, 
introduced in 1931 by Woodruff and Goodpasture (685), is the fer- 
tilized chicken egg. It is no exaggeration to say that the use of chick 
embryos has opened a new era in the study of animal viruses for both 
theoretical and practical purposes (77). If anything, one may regret 
that the quick realization of the great potentialities of egg-culture 
methods has for many years caused a certain neglect of the more 
refined types of tissue cultures, which can contribute much to the 
analysis of virus growth in the host cell. 

The chick embryo offers numerous advantages as an experimental 
host in virus work: ready availability and low cost; absence of latent 
viruses (see, however, 107); absence of antibodies; presence of a 
variety of extraembryonal organs, well-known from embryological stud- 
ies and suitable for the study of the behavior of many viruses in single 
layers of cells. 

Embryological notes. 1 When a fertilized hen egg is incubated at 
38-39 C, the embryo develops and the chick hatches after 21 days. 
The embryo develops from the fertilized egg nucleus, which is situated 
between yolk and albumen under the vitelline membrane. By cleav- 
age, the fertilized nucleus gives rise to a primitive disc of cells. This 
disc becomes differentiated into an inner cell layer, the entoderm, and 
an outer layer or ectoderm. Between these a thickening appears in the 
central area (primitive streak), due to a deep crowding of cells that 
give rise to the middle layer or mesoderm ( figure 82 ) . The three primi- 
tive layers ectoderm, mesoderm, and entoderm give rise to all tissues 

1 The student should familiarize himself more thoroughly with this subject by 
study of a standard textbook of embryology. Most such textbooks have a chapter 
devoted to the development of the chick embryo. 


240 Viruses in the Chick Embryo 

and organs of the body. The ectoderm gives rise mainly to the skin, 
the central nervous system, and some endocrine glands; the mesoderm 
gives rise to connective and muscular tissues and to the circulatory 
system; the entoderm, to the intestinal mucosa and its related secretory 
glands. In addition, the primitive layers develop beyond the area of 
the body of the embryo proper and form the embryonal membranes, 
which surround the embryo in the egg and which are of great interest 
to virus workers. 

The mesoderm soon splits into two layers separated by the coelomic 
cavity. While the development of the embryo proceeds, the membrane 
constituted jointly by entoderm and mesoderm (spanchnopleure) ex- 
tends around the yolk and, inside the embryo, lines the intestine. The 
membrane constituted jointly by the ectoderm and the outer mesoderm 
(somatopleure) rises in folds all around the embryo body proper. As 
the folds close over the embryo, they form a sac, the amniotic cavity, 
in which the embryo becomes suspended. The folds dig deeper and 
deeper around the body of the embryo, leaving only a narrow opening 
on the ventral side the primitive umbilicus through which the intes- 
tine of the embryo communicates with the yolk sac, its main source of 
food. The closure of the amniotic sac over the embryo separates the 
amnion from the- outer part of the ectoderm, which lines most of the 
inner surface of the egg immediately under the shell membrane and 
forms the serosa or chorion (figure 82). 

Late in the third day of incubation, the entoderm from the hind part 
of the primitive intestine gives rise to a new sac, the allantois, which 
pushes its way out of the body of the embryo, between the yolk sac 
(now surrounded by the spanchnopleure) and the amnion. The al- 
lantois ultimately fills the whole space under the chorion. Inside the 
allantoic cavity a clear fluid is secreted. As a result of the increasing 
size of the allantoic cavity, the embryo at the age of 8 to 12 days, when 
it is generally used for virus work, presents the following structure. 
Under the shell we find the shell membrane; at the wide end of the egg 
the shell membrane is divided into two layers, which delimit the "air 
space." Under the shell membrane there is a layer of epithelium, 
representing the ectodermal layer of the chorion. Beneath this we find 
a mesodermal layer derived from the fusion of the mesoderm of the 
chorion with the mesoderm surrounding the allantoic sac. Next, we 
find the entodermal epithelium lining the allantoic cavity. Together, 
the chorion and the external lining of the allantoic cavity form the 
chorioallantoic membrane. 

CH, 12 Viruses in the Chick Embryo 241 

Across the allantoic cavity, which is filled with allantoic fluid, the 
allantoic epithelium rests over a mesodermal layer, which lies in con- 
tact with the amnion and the yolk sac. The amnion itself is lined 
hy an ectodermal epithelium, in direct continuity with the skin of the 
embryo around the umbilicus, and contains a limpid fluid. After the 
twelfth day of incubation, the amniotic fluid begins to receive some of 
the albumen through a newly established communication between the 
egg white and the amnion. The yolk sac is lined by an entodermal 

The mesoderm surrounding the allantoic cavity is rich in blood 
vessels, which receive blood from the embryo, circulate it into a rich 
network of capillaries, and return it to the body. The allantoic mem- 
brane constitutes the respiratory organ of the embryo; gaseous ex- 
changes take place between the blood in the capillaries and the atmos- 
phere, separated only by the thin chorionic epithelium, the shell 
membrane, and the porous egg shell. The presence of blood-filled 
allantoic vessels, easily seen by "candling" a fertilized egg, represents 
the best sign of viability of the embryo. The blood vessels become 
indistinct soon after circulation stops. 

Techniques in virus work (77; 655). All parts of the chick embryo 
have been used for growing viruses: the allantoic cavity, the amniotic 
cavity, the yolk sac, and various organs of the embryo itself. Growth 
of the viruses takes place in the cells that line the various cavities. 
Inocula should, of course, be bacteria-free. If the material to be inocu- 
lated cannot be made bacteria-free (as, for example, in first isolations 
of some viruses from throat washings, whose filtration might remove 
the virus) bacterial infection may be prevented by the addition of 
antibiotics. Eggs are incubated at about 38-39 C until they reach 
the desired age. They are candled -before use to check their viability 
and to mark on the shell certain structures, such as the limits of the 
air sac, as a guide in the following operations. 

Inoculation into the allantoic cavity can be done in two ways. After 
drilling a small hole over a well-vascularized area in the shell, we can 
introduce the inoculum with a needle or a fine pipette through the 
shell membrane and the chorioallantoic membrane. Or we can drill 
a hole over the air sac near its edge and introduce a needle into the 
allantoic cavity through the hole. In drilling the egg shell over the 
allantois, it is important to avoid damaging the chorioallantoic mem- 
brane; hemorrhage in the membrane often results in death of the 


Viruses in the Chick Embryo 

Inoculation can be made onto the outer surface of the chorioallantoic 
membrane. First, a triangle or rectangle of shell is drilled and lifted 

lateral amniotic fold 
lateral body fold 






ectoderm \ 

somato . 

s t )splanch- 



amniotic cavity 



amniotic fold 

( somatopleurc 
serosa ' 

allantois / splanchnopleure ) 
yolk stalk 

nopleure \ 


k vitelline membrane 

Figure 82. The development of the chick embryo. A, Early in the second day 
of incubation. B, Early in the third day of incubation. C, Fifth day. D, Ninth 
day. From: Patten: Embryology of the Chick, 4th Ed., Blakiston, New York. 

out. A small drop of saline solution is deposited on the shell mem- 
brane, which is then gently slit with a blunt needle. Suction, applied 
through a small hole drilled over the air sac, causes the chorioallantoic 

CH. 12 

Viruses in the Chick Embryo 


membrane to drop, that is, to pull away from the shell membrane. 
This creates a false air space, into which the inoculum can be deposited. 



allantoic cavity 



allantoic cavity 




Figure 82 (continued). 

( splanchno- 
pleiire ) 






belly stalk 

For better visualization, the exposed shell membrane can be removed 
and replaced by a sterile cover slip, which is sealed to the shell with 
paraffin (figure 83). 

Amniotic inoculation can be done either by reaching the amnion 
through a false air space, or, more speedily but with lower chance of 


Viruses in the Chick Embryo 

;uccess, by sending a needle through the natural air sac toward the 
embryo's body while candling the egg. When the tip of the needle 
touches the embryo, the needle's opening is in the amniotic cavity. 

Yolk-sac inoculation, for most routine purposes, is done by intro- 
lucing the needle to a depth of about 1 to V/ 2 inches vertically into 

Figure 83. False air space for inoculation of the chorioallantoic membrane of 
a chick embryo. From: Buddingh. In: Rivers, Viral and Rickettsial Infections of 
Man, Lippincott, Philadelphia. Courtesy Dr. G. J. Buddingh, Louisiana State 
University, New Orleans. 

the blunt end of the egg. Intravenous inoculation and direct intro- 
duction into various organs of the embryo have also been described. 

After inoculation, the eggs are returned to an incubator (generally 
at a temperature of around 35 C, more favorable for several viruses 
than 39 C) and examined after a period of time determined by ex- 
perience, generally 2 to 4 days. The presence, growth, and effects of 
viruses are detected by a variety of criteria, depending on the virus 
and on the route of inoculation. Death of the embryo occurs regu- 
larly with such viruses as equine encephalomyelitis or Newcastle dis- 
ease. In some diseases, specific pathological lesions appear in the body 
of the embryo or in the extraembryonal membranes. An increase in 

CH. 12 Virus Culture in Eggs 245 

virus titer either in the extraembryonal fluids or in ground embryonal 
or extraembryonal tissues is a valid sign of infection. Serological 
tests on the allantoie or amniotic fluids or on emulsions of the yolk sac 
or of the whole embryo may be used to demonstrate virus antigens. 
In some cases, elementary bodies may be stained and counted in the 
infected fluids. 


A majority of the animal viruses can reproduce in the chick embryo. 
Reports on cultivation of trachoma and dengue viruses are doubtful. 
Some viruses will grow only if inoculated by a given route, for 
example, the rabies virus only if injected into the brain or on the 
chorioallantoic membrane, from which it spreads to the central nervous 
system (72). Successful inoculation may depend on the age of the 
egg. Influenza virus inoculated in the amniotic cavity gives a higher 
proportion of positive inoculations in 13-day-old eggs than in younger 
ones (329); this has been attributed to the active swallowing of 
amniotic fluid by the older embryo, which brings the virus in contact 
with the lungs. Thus, the older eggs are used for diagnostic inocu- 
lation of throat washings, where the goal is to obtain the highest 
possible proportion of positive results. Other viruses, for example, 
mumps (670), grow better or at least reach higher titers in younger 

Living cells are necessary for virus growth, although the embryo 
itself need not be viable. Influenza virus has been grown in the 
allantoie cavity of eggs previpusly kept for several days in the re- 
frigerator and returned to 35 C; the lining cells are still viable, as 
shown by their continuing metabolism and by their ability to repro- 
duce if transferred to tissue cultures. Chilling the eggs at 30 C, 
a treatment that kills all cells, also suppresses their ability to support 
virus reproduction (394). 

Little is known of the mechanisms that underlie the specific ability 
of the cells that line different cavities in the chicken egg to act as hosts 
for given viruses. Certain regularities are apparent; for example, the 
viruses of the psittacosis-lymphogranuloma group grow well in the 
yolk sac, and so do rickettsiae. The fact that the ability of a virus to 
grow in various locations varies with the age of the embryo suggests 
that developmental changes in physiological conditions are of im- 
portance. It should be remembered that the conclusion as to whether 
a virus grows or not has generally been reached from tests of recovery 

246 Viruses in the Chick Embryo 

of virus in the various fluids. This must depend in part on the release 
of virus from the cells and not only on its ability to grow. 

A number of viruses retain their tissue specificity in the chick em- 
bryo, even though they are able to grow in the undifferentiated cells 
of the extraembryonal membranes. Thus, the Rous sarcoma virus 
infects both ectodermal and mesodermal cells of the chorioallantoic 
membrane, but tumors only arise through proliferation of mesodermal 
cells (369). The rabies virus maintains its affinity for the central 
nervous system (72), whereas influenza viruses spread from the chorio- 
allantoic membrane to give typical lung diseases. 

New affinities, of great practical and theoretical interest, are revealed 
in the chick embryo. Thus, influenza virus A, inoculated into 2-day-old 
eggs under the vitelline membrane directly over the embryos, kills 
them with typical developmental abnormalities which affect mainly 
the central nervous system; 4-day embryos are also killed, but no evi- 
dent brain malformations appear (289). These observations, on the 
one hand, point to a potential neurotropism of the influenza virus, 
which has recently been confirmed by other observations (582). On 
the other hand, they illustrate the role of developmental changes in 
determining the pattern of infection. A striking case in man is that 
of the virus of German measles, which causes a mild disease in children 
or adults, but produces severe birth defects in children of mothers who 
have contracted German measles in early pregnancy (277). 

Goodpasture and his collaborators have introduced an elegant re- 
finement for the study of virus specificity in the chick embryo. This 
consists in grafting pieces of foreign tissues on the chorioallantoic 
membrane, which is then inoculated with the viruses. In a series of 
experiments (261), the epithelium from human skin or human placenta 
was grafted and inoculated either with fowl pox virus or with a typical 
mammalian virus, the causative agent of mare abortion. The first virus 
failed to grow on the human graft, but gave typical lesions on the 
surrounding membrane, whereas the mammalian virus only grew on 
the graft. Another experiment (260) was made to test whether im- 
munity to fowl pox, acquired as a result of a previous infection, was 
due to immunity of the individual cells or to humoral mechanisms. 
The skin from either susceptible or immunized chickens was grafted 
onto the chorioallantoic membrane, and fowl pox virus was added; 
both grafts became infected with equal ease. A technique has been 
described that permits the transplantation of whole corneas from man 
or rabbit onto the chorioallantoic membrane for study of viruses with 
strictly mammalian host range (374). 

CH. 12 Growth of Viruses 247 


The embryonated egg has provided a convenient host system for the 
study of the process of reproduction of animal viruses. Although these 
studies have not yet reached the stage of detailed analysis of virus 
growth obtained with bacteriophage, they have provided much infor- 
mation of value. 

Cytologically, the growth process has been followed by observations 
of embryonal membranes at intervals after infection. The chorio- 
allantoic membrane, inoculated through a false air space, has been 
used mainly for vaccinia and other mammalian pox viruses (321; 322); 
and the yolk sac has been used for the psittacosis-lymphogranuloma 
group (544). Moreover, extensive studies on the histology of early 
infection have been made on the chorioallantoic membrane with a 
number of viruses (see 77). 

This membrane, after infection with vaccinia or mouse-pox (ectro- 
melia), was observed directly in vivo by the ingenious device of 
lifting the membrane against a sterile cover slip sealed onto a window 
in the shell over the false air space. When pressure is applied through 
a hole in the air-sac shell, the membrane rises against the cover slip 
and can be observed microscopically under oblique illumination from 
above ("ultropak" system; 322). The infected cells show elementary 
bodies undergoing a series of modifications, similar to those observed, 
for example, in infected cells of the rabbit cornea. The resulting in- 
clusions resemble typical Guarnieri bodies. Results of the same type 
have been described from histological observations on fixed chorio- 
allantoic membranes (321). 

A detailed study of the histology of yolk sac infected with lympho- 
granuloma venereum (544) showed a more complex cycle, with en- 
larged elementary bodies giving rise to large homogeneous bodies or 
"plaques." The plaques are later transformed again into several ele- 
mentary bodies, which escape and infect other cells. This process is 
very similar to observations on the related viruses of the psittacosis 
group in tissue cultures or in the corneal epithelium (89; 418; see page 

Quantitative analyses of growth in egg cultures have been published 
by various authors, and the influence of a number of factors on rate 
and total amount of virus increase have been studied. For example, 
growth curves have been published for influenza and for mumps 
viruses in the allantoic cavity, showing the faster rate of growth of 

248 Viruses in the Chick Embryo 

the influenza virus (10 G -fold increase in 18 hours instead of several 
days; 461; 670). The effect of a variety of chemicals on virus growth 
in the allantois has been described, and screening of antibiotics against 
viruses by such technique has become a routine procedure. The prac- 
tical results, however, have been limited, and thie main successes have 
been scored in the study of the growth cycle of influenza viruses in 
the allantoic cavity and of the effect of various agents on this cycle. 

Since these studies have derived their main impetus and in part also 
their interpretation from the discovery and analysis of the peculiar 
phenomenon of the agglutination of red blood cells by certain viruses 
( hemagglutination ) , we shall defer their discussion to chapter 13, in 
which we describe the hemagglutination phenomenon. 


Allantoic inoculation, and in some cases yolk-sac inoculation, are 
commonly used to obtain certain viruses in large amounts for vaccine 
production. Virus produced in this way is free of bacterial contami- 
nations. Before it can be used as a vaccine, it must be separated from 
components of the allantoic fluid that may cause allergic reactions in 
the vaccinated subjects. Partial purification of the virus can be done 
by differential centrifugation, since the normal allantoic fluid has been 
found not to contain any component that sediments as fast as most 
viruses; the largest normal components of allantoic fluid have sizes of 
the order of 30 to 40 ny. In viruses of the influenza group, purification 
is easily performed by letting the virus be adsorbed in the cold by red 
cells from the allantoic blood vessels, then causing the virus to be 
eluted from the cells by incubation at 37 C (see chapter 13). Yellow 
fever vaccine is a desiccated homogenate of whole embryos inoculated 
at 7 days of age with strain 17D and incubated for 4 days (292). 

Embryonated eggs are frequently used for the direct isolation of 
viruses from patients, as a diagnostic procedure. Some viruses will 
grow best in the allantoic cavity, others in the amniotic cavity or on 
the chorioallantoic membrane. It is important to remember, however, 
that the affinities of a virus for different organs of the embryo often 
vary upon subculture, so that the route of inoculation to be used for 
primary isolation from a patient is often different from the one that 
gives the best growth response with egg-adapted viruses. For example, 
influenza virus, according to several authors, can best be detected by 
direct inoculation of washings from the human throat into the am- 
niotic cavity, although it will later grow as well or better by allantoic 

CH. 12 Practical Applications of Egg Cultures 249 

inoculation (329). This is probably due to modifications in the 
growth habits of the virus in the course of egg passage; variants be- 
come established, which grow better than the original type in the new 

Radioactive influenza virus, containing P 32 , was obtained by intro- 
ducing labeled phosphate in the allantoic cavity either before or shortly 
after inoculation with the virus ( 266 ) . 

The chorioallantoic membrane has been used for the titration of 
several viruses (77). Small aliquots of suspensions of viruses such as 
vaccinia, herpes, ectromelia, fowl pox, myxoma, laryngotracheitis, and 
many others cause the formation of discrete, localized lesions or pocks, 
whose appearance, both macroscopic and microscopic, is more or less 
characteristic for each virus infection (figure 3). Under controlled 
conditions of inoculation, the number of these lesions is a linear func- 
tion of the amount of virus inoculated, provided the amount of virus 
is not so great that the lesions become confluent. For different viruses 
the maximum countable number of lesions varies between 20 and 100. 
Titers obtained by this method are in good agreement with titers ob- 
tained by end-point methods. The chorioallantoic titration has been 
used in determining virus survival in the titration of virus-neutralizing 
antibodies. Mixtures of virus and antibody are compared with virus 
alone for the number of pocks produced, and the neutralizing potency 
of the antibody is estimated. 

The allantoic sac inoculation provides a convenient method for 
titration of certain viruses by the end-point method, for example, in 
the measurement of the neutralizing power of antibodies and of the 
effect of drugs on virus reproduction. The presence of infectious virus 
can be detected as an all-or-none response from the presence or absence 
of large amounts of virus after 2 or 3 days of incubation, provided the 
new virus can be detected directly by some in vitro test. This is true, 
for example, of viruses that agglutinate red blood cells or that fix com- 
plement readily. This procedure eliminates the need for animal inocu- 
lation in detecting the virus. Thus, by replacing the use of animals 
with that of eggs, such procedures have greatly reduced the cost of 
virus work. 



Hemagglutination Phenomena 

and Virus Growth 
Summary on Virus Reproduction 


A fortunate discovery by Hirst in 1941 (323; 324) has made possible 
a remarkable progress in the study of the relation of certain animal 
viruses to animal cells. In harvesting the allantoic fluid of eggs in- 
fected with influenza virus, Hirst observed that if blood was shed into 
virus-containing fluid the erythrocytes quickly became agglutinated. 
No such agglutination occurred in uninfected fluids. The agent re- 
sponsible for this "hemagglutination" is the virus particle itself. The 
particles are taken up by the red blood cells; this can be shown either 
by titration of free virus or by observation in the electron microscope 
of virus adsorbed on the red blood cells (296). Antivirus serum in- 
hibits hemagglutination. The inhibitory titer of a serum is related, 
although not strictly proportional, to its virus-neutralizing titer. 

Hemagglutination is detected either by an increase in the rate of 
sedimentation of a suspension of red blood cells measured by the 
residual turbidity of the supernatant (figure 84), or by the so-called 
"pattern test" (575). Nonagglutinated cells sediment into a round, 
solid pellet at the bottom of a test tube, whereas agglutinated cells 
give rise to a characteristic, scattered pattern of sedimentation (figure 
85). A hemagglutination test can be read within an hour or less and 
represents a convenient method for the detection of viruses with the 
hemagglutinating property. 

Hemagglutination has been observed with three groups of viruses. 
One group includes the influenza, mumps, and Newcastle disease 
viruses. A second group consists of vaccinia, smallpox, and ectromelia 
viruses. A third group includes a variety of other viruses: Japanese, 


CH. 13 Hemagglutination by Viruses 251 

St. Louis, and Russian encephalitis; mouse encephalomyelitis (Theiler's 
GD-VII); West Nile fever; encephalomyocarditis group; pneumonia 
virus of mice and fowl plague. 

Different groups of viruses cause different types of agglutination. 
Those of the vaccinia group ( vaccinia-variola-ectromelia ) cause an 
agglutination that is readily reversible by heating, by treatment with 
antivirus serum, and by changes in pH and in salt concentration. The 
properties of the resuspended red blood cells are unchanged and the 

Figure 84. Hemagglutination by influenza virus. Successive 1:2 dilutions of 
virus. Complete agglutination in tubes 1-5. From: Research Today, 8:26, 1952. 
Courtesy Dr. O. K. Behrens, Lilly Research Laboratories, Indianapolis. 

cells are reagglutinable (133). The vaccinia hemagglutinin can be 
separated from the virus particles themselves, and appears to combine 
in a rather unspecific way with the red cell surface. Red cells from 
only about 50% of chickens are agglutinated. Red cells from most other 
animals are not. 

The agglutination produced by viruses belonging to the so-called 
MNI group (mumps, Newcastle, and influenza) is more interesting 
from a biological standpoint. In the following section we shall mainly 
be concerned with the viruses of this group. Hemagglutination by 
most viruses of the third group has several similarities with that pro- 
duced by the MNI group, but without elution phenomena or any other 
signs of enzymatic activity by the viruses. 

The MNI viruses agglutinate red blood cells of a variety of different 
animals, for example, chicken, guinea pig, and man. The correlation 
between the ability of an animal to act as host for a virus and the 


Hemagglutination Phenomena 

CH. 13 Hemagglutination by Viruses 253 

agglutinability of its red blood cells by that virus is good for some 
viruses, for example, pneumonia virus of mice, but poor for most others. 
Moreover, there are variations among red blood cells of individual 
animals and among strains of the same virus. For example, strains of 
influenza A, when first isolated from man, are in the O (original) form; 
upon transfers in the chick embryo a variant form, D, appears. This 
differs from the O form in giving a higher ratio between the ag- 
glutinating titer for chicken red cells and that for guinea pig red cells 
(116; 118). 

The optimum virus concentration for agglutination of red blood cells 
(at a concentration of approximately 100,000 cells per ml) is about 
10 4 -10 5 virus particles per cell. With purified influenza virus, it has 
been calculated that hemagglutination can occur with equal numbers 
of cells and virus particles (242). Electron micrographs suggest that 
the agglutination is due to the formation of a lattice, in which red blood 
cells are held together by multivalent virus particles (296). Hemag- 
glutination provides a rapid titration method, which is 100,000 to 
1,000,000 times less sensitive than titration of infectivity, but which is 
advantageous in a variety of tests. 

Elution of viruses. The adsorption of virus particles to red cells is 
not necessarily permanent. If a virus-cell mixture, in which aggluti- 
nation has occurred, is incubated at 37 C, the virus is soon released 
more or less completely. The elution of virus is slow or absent at 
C. The red cells that have released an excess of virus become 
inagglutinable either by the same virus or by newly added virus of 
the same strain. The eluted virus, however, is unchanged and capable 
of agglutinating other cells; the agglutination and release cycle can 
be repeated indefinitely with successive batches of cells (324). 

These observations were interpreted by Hirst as indicating that the 
release phenomenon resulted from an enzymatic activity of the virus 
particles on a substrate located on the surface of the red cells and 
needed for stable combination. Thus, the virus particles combine with 
specific receptors on the cell surface, destroy enzymatically the re- 
ceptors or some essential portion of them, and render the cells non- 
agglutinable. All evidence accumulated since Hirst's discovery has 
confirmed this viewpoint. 

The changes produced by the virus are evidenced by a variety of 
alterations in the red blood cells. New antigenically reactive groups 
are revealed; these are the T-agglutinogens, responsible for the so- 
called "panagglutinability" of such cells by a variety of sera (115). 
The electrophoretic mobility of the red cells is altered in a way that 

254 Hemagglutination Phenomena 

is more or less specific for each virus (631). Electrophoretic tests with 
red blood cells, whose receptors have been destroyed by various viruses 
of the MNI group, suggest that the respective receptors partially over- 
lap. Hemolysis has been observed with mumps and Newcastle dis- 
ease virus (119; 132); cells from which receptors for hemagglutination 
have been removed are not lysed by the viruses. 

Definite evidence for the relation among red blood cell receptors for 
various MNI viruses is found in studies of the agglutinability of cells 
from which a virus has been eluted. Elution of mumps virus leaves 
the cells still agglutinable by Newcastle and influenza; elution of New- 
castle leaves the cells agglutinable by influenza only; influenza A leaves 
red cells still agglutinable by some B strains, and so on. On this basis 
a receptor gradient (see 115) can be constructed (table 24). The 
viruses are supposed to destroy a certain fraction of a population of 
receptors possessing different, specific affinities for other viruses. Each 
virus would appear to destroy its own receptors, which include the 
receptors for the viruses that precede it in the receptor gradient scale. 

This picture, however, is probably an oversimplification. In fact, 
every virus of the MNI group is capable of completely destroying the 
receptors for all other viruses, provided a sufficient amount of virus 
is allowed to act for a long enough time (330). The most likely ex- 
planation is tMt the different viruses are adsorbed, with different 
affinities, by common or overlapping receptor areas, which contain 
substrates for the virus enzymes (332). Some viruses, once attached, 
may not elute until they have digested all or most of the substrate ( and 
the receptor ability with it), whereas others may leave behind enough 
undigested substrate to permit virus strains possessing greater affinity 
to combine with the receptor. Virus 1233 (influenza virus C), how- 
ever, apparently has an independent system of receptors in the red 
cells of some animals. No mutual destruction of chicken red cell 
receptors is observed between influenza C and the viruses of the MNI 
group (331). 

Hemagglutination and the relation of viruses to host cells. The 
discovery that a group of viruses possessed an enzymatic activity for 
substrates located on the surfaces of red blood cells was enough in 
itself to stimulate the analysis of the enzymes and of the substrates 
involved, even though the red blood cells are not host cells, in the sense 
that they cannot support virus reproduction. Another discovery by 
Hirst, however, indicated that the hemagglutination phenomenon had 
a direct bearing on the process of virus infection (325). Excised lungs 
from ferrets were perfused to eliminate blood, then filled through the 

CH. 13 Hemagglutination and Virus-Host Relation 255 






1 1 1 1 1 1 1 



" S 

+ 4- + + + + 1 


w H 

+ + + + + + 1 






+ + + + + 4- 1 









-f + 4-4- -^ 1 1 

* j 




| 1 


o J, 


4-4-4-4-1 1 1 


t> 5" 

^ CM 


01 H 

4- + 4- 4- 1 1 I 

S 3 







a 2 

J: S 




00 j 

4- 4- -H I 


O . 

S ^ 



1 *"" 

C cq 

^ ^ 

4- 4" 4- 1 1 1 1 

.2 10" 

irt c" 



1 "*! 


W) T3 



CO >-J 

4- 41 1 1 

"3 W as 




2 g 



1- 5 

^ > 

4- 4- 1 1 1 1 1 


o ~ 

3 i s 

JJ ^ 



E! I 


4- III 

o J 2 

-G CO* 


M W 

4- 41 I I I l I 




| S 1 



2 H | 


<N Q 

4- 1 1 1 1 1 1 

"e <N 



4- 1 1 1 1 1 1 



>> ** 

c .5 c 





J3 -G 

1 S ^ 

rl 10 

.S H " 



Q Q S S 

3 .- ^J 


. Q M 55 "-I {" f 

"8 a a 




co s HH PQ cc 02 


N o r>. o M 

256 Hemagglutination Phenomena 

bronchi with influenza virus suspensions. The virus was quickly taken 
up by the cells of the lungs and then, upon incubation at 37 C, became 
eluted. If the lungs of a living ferret were flooded with virus, the virus 
was likewise taken up, but soon became nonelutable and infection 
followed. Similar experiments have been made, with identical results, 
using mouse lung or allantoic membrane of chick embryo. 

These observations indicated the similarity between the adsorption 
of influenza viruses onto red blood cells and onto host cells. Hirst 
suggested that the enzymatic activity of the virus on the host cell 
receptors may be a fundamental stage of the virus attack on the sus- 
ceptible host cell. Two questions, thus, became preeminent. First, 
what are the properties of the virus enzyme and of its cellular sub- 
strates? Second, can this knowledge be used to investigate the 
mechanism of cell-virus interaction, and possibly to devise means that 
would prevent infection? With such perspectives, work on hemag- 
glutination and allied phenomena has progressed rapidly. 

A variety of treatments, when properly administered, preserve the 
agglutinating ability of the virus particles but suppress their ability 
to elute from red blood cells (enzymatic activity). Heat, ultraviolet 
light, and various chemicals act in this way ( 327 ) . The eluting ability 
is more resistant to various treatments, such as formalin or ultraviolet, 
than the reproductive ability of the virus, but is much less resistant 
than the agglutinating ability. 

The viruses of the MNI group differ in their resistance to various 
agents. It has been widely maintained that the suppression of eluting 
ability results from inactivation of the enzymatic function without 
suppression of the affinity of the enzyme for its substrate, so that a 
stable combination results. An alternate possibility is that the com- 
bination between virus and cell is mediated by chemical configurations 
that are different or more inclusive than the enzyme itself; the enzyme, 
when active, would act to destroy the cell receptors without itself being 
fully responsible for attachment to the receptors. Evidently, more 
information about the nature of enzyme, substrate, and receptors is 
needed to provide an answer to these questions. 

The receptor-destroying enzyme. The timely introduction of a 
model system has helped further progress. Having noticed the simi- 
larities between red cells modified by virus elution and red cells altered 
by certain bacterial filtrates, particularly from Vibrio cholerae, Burnet 
and his coworkers succeeded in demonstrating in such filtrates a "re- 
ceptor-destroying enzyme" ( RDE ) , which duplicates most of the virus 
effects on red blood cells (see 115). Exposure of red blood cells to 

CH. 13 Hemagglutination and Virus-Host Relation 257 

RDE progressively destroys their susceptibility to agglutination by 
viruses of the MNI group. The order in which agglutinability by dif- 
ferent viruses is lost closely parallels the position of the viruses in the 
receptor gradient; agglutinability by mumps is lost more readily than 
that by Newcastle virus, and so on. 

The RDE, as well as active virus (328), can remove from red cells 
virus particles that have become unable to elute themselves. This 
suggests a more complex basis for virus-cell combination than a direct, 
1:1 combination between enzyme and substrate molecules. Since 
another enzyme (RDE or active virus) can reverse the combination, 
it seems possible that the cell receptor contains some substrate groups, 
uncombined with the virus enzyme and available to other enzymes, 
which are operative in holding virus and cell together. RDE can 
remove virus receptors from the cells of host tissues, such as mouse 
lung or allantoic sac epithelium. The allantoic membrane, after treat- 
ment with sufficient RDE, remains unable to adsorb virus for as long 
as 24 hours; if the RDE is removed from the allantoic cavity, the 
epithelium can regenerate its adsorbing capacity (630). Even active 
virus already attached to host cells may still be removed by RDE if 
the RDE is introduced soon after the virus. Trypsin inactivates RDE 
and reportedly also the virus enzyme. The receptor-destroying ac- 
tivity of RDE requires calcium ions and is suppressed by citrate, a 
behavior that is shared by the enzyme activity of some of the MNI 

Virus substrates. The RDE provides a convenient model system for 
virus enzymes. The next question concerns the substrates for these 
enzymes. A suitable method for the detection of virus substrates is 
the use of agglutination-inhibition tests. 

Heat treatment or other manipulations can, as we have seen, render 
a virus agglutinating but not elutable. It was observed that a variety 
of animal fluids and tissue extracts-including urine (642), blood 
serum (230), allantoic fluid (636), and egg white (398) -can inhibit 
hemagglutination by the nonelutable virus. Hemagglutination by un- 
modified virus is inhibited to a very slight extent, if at all. The 
normal virus, supposedly, is not subject to inhibition because it can 
destroy the inhibitors enzymatically, whereas the modified virus, which 
has lost its enzymatic activity, combines with the inhibitors and is 
thereby prevented from combining with red blood cells. Thus, the 
inhibitors are revealed by their ability to prevent hemagglutination by 
modified virus, which acts as indicator for the inhibitors. 

258 Hemagglutination Phenomena 

A variety of inhibitors has been reported. Some occur free, like the 
so-called Francis inhibitor in blood serum (230); others can be ex- 
tracted from red blood cells (164) or other cells. A common property 
of all these inhibitors appears to be their content of polysaccharides. 
Their activity is suppressed by reagents such as periodates, which are 
known to affect polysaccharides fairly specifically. It is now generally 
believed that most inhibitors belong to the class of mucins or muco- 
proteins, a rather ill-defined group of proteins conjugated with carbo- 
hydrates (16). Most materials rich in mucin, such as the fluid from 
ovarian cysts and from mucous secretions, exhibit inhibitory activity. 
Usually the inhibitory activity can be destroyed not only by active 
virus but also by preparations of RDE. The most thoroughly analyzed 
inhibitor, that from egg white, has been partially purified and obtained 
in the form of a highly viscous material, consisting of filamentous par- 
ticles, about 8 X 10 mol. wt. (400). 

It seems altogether reasonable to view the combination between 
virus and cell receptors and that between virus and inhibitors as 
enzyme-substrate combinations. The inhibitors compete with the cell 
receptors for the virus enzymes, which, if active, can digest them. 
This viewpoint can account for most of the known facts. Supplemen- 
tary hypotheses will probably be necessary to account for specific 
phenomena observed with modified viruses and with various inhibitors. 
A kinetic analysis of the inhibition of heated swine influenza virus by 
egg-white inhibitor (399) has led to a rather simple picture of com- 
petition between inhibitor and red blood cells for the active groups on 
bivalent virus particles, the combinations being practically irreversible. 
Apparently, each collision between virus and inhibitor or red blood 
cell is effective in producing combination. This suggests that electro- 
static forces play a role in the attachment reaction. 

Not all inhibitors are related to the cell receptors for MNI viruses. 
Thus, allantoic fluid contains two inhibitors: one is not destroyed by 
virus, is probably lipidic in nature, and acts in a nonspecific way; the 
other is a specific inhibitor, which is destroyed by virus (636). 

According to some workers (15), the picture of virus-receptor and 
virus-inhibitor combinations as enzyme-substrate combinations is an 
oversimplification. One objection is that hemagglutination by heated, 
enzymatically inactive influenza virus is more readily inhabited than 
the hemagglutination by active virus in the cold, where enzyme ac- 
tivity should be absent. The difference seems hard to explain, if 
inhibition is simply an enzyme substrate combination without enzyme 

CH. 13 Hemagglutination and Virus-Host Relation 259 

activity. Since the enzymes are somewhat active even at C, how- 
ever, this argument is debatable. 

Another objection is that a brief treatment of certain inhibitors with 
periodate increases their inhibitory power; it is pointed out that 
chemical treatment should not increase the affinity of a substrate for 
an enzyme. It is possible, however, that the combination between 
indicator virus and inhibitors is a reversible reaction, which reaches 
a finite equilibrium ( 291 ) . Periodate treatment of the inhibitor might 
simply affect the equilibrium value. 

Virus enzymes and infection. The fact that cells which support 
growth of MNI viruses possess receptors digestible by the viruses or 
by RDE raises the question of the role of the virus enzymes in infec- 
tion. Destruction of the receptors in lung or allantoic cells during in- 
fection with influenza viruses is the rule, and it is an appealing hy- 
pothesis that the virus, following its uptake by the surface receptors, 
works its way into the cells by digesting the receptors (332). 

It must be admitted that the evidence for this hypothesis is indirect; 
no proof exists of the role of virus enzyme in penetration. It is possible 
that the enzyme's main function is the destruction of mucous films on 
the surface of susceptible tissues. 

The recognition of mucoproteins as virus inhibitors and receptors 
has suggested the use of polysaccharides as . competitive inhibitors, 
which may suppress infection by competing with cell receptors for 
the virus. Indeed, a number of substances, including pectins, bacterial 
polysaccharides, and others, interfere both with hemagglutination and 
with the growth of some of the MNI viruses (273). The capsular 
polysaccharide from Klebsiella pneumoniae ( Friedlander's bacillus) is 
an example (256'). It inhibits infection with mumps or with pneu- 
monia virus of mice, two viruses which interfere with one another in 
mixed infection, probably at the intracellular level, and which are 
characterized by rather slow reproductive cycles ( 252 ) . The growth- 
inhibiting power is still present if the polysaccharide is introduced as 
late as 4 days after virus inoculation. It seems possible that this poly- 
saccharide blocks infection by an intracellular action rather than by 
interference with virus adsorption. If this is correct, the false lead 
derived from the study of inhibitors of hemagglutination may have led 
to a potentially fruitful new approach to antiviral therapy. 

Inhibitors and virus purity. The allantoic fluid contains inhibitors 
of viruses of the MNI group. Not all this inhibitory power is de- 
stroyed by active virus. Indeed, infected allantoic fluids contain not 
only free virus but also some in combination with inhibitors (291). 

260 Hemagglutination Phenomena 

This situation must affect the apparent chemical composition of virus 
isolated from the allantoic fluid. This is of interest in connection with 
the cross-reaction that influenza viruses grown in eggs give with anti- 
serum against normal allantoic fluid, a cross-reaction which has been 
noted repeatedly (377). This cross-reaction probably reflects the 
presence of inhibitors or other host materials attached to the virus 
particles rather than the incorporation of host antigens into the virus 
protoplasm itself. 

A more complex situation is encountered with pneumonia virus of 
mice (349). Virus extracted by grinding infected mouse lung is in- 
fectious but does not agglutinate red blood cells. Brief heating of the 
virus suspension at 80 C renders the virus noninfectious, but confers 
upon it a hemagglutinating ability for red blood cells from mouse or 
hamster (not from other species). Addition of ground, unheated, 
normal lung material again suppresses the hemagglutinating power. 
Apparently, in the ground lungs the virus is combined with a com- 
ponent of the host tissue, which, without suppressing infectivity, pre- 
vents the adsorption of virus on red blood cells. Virus can be ob- 
tained free of this inhibitor, either by collecting the fluid that seeps out 
of a sharp cut in the infected lungs, thereby avoiding the grinding of 
lung cells, or by dissociating the virus-inhibitor combination at low 
electrolyte concentration. The virus thus obtained is highly infectious, 
agglutinates red blood cells, and consists of particles about 30-40 m/x 
in diameter, whereas those obtained by grinding average 140 m/x. 

In this case the inhibitor is not destroyed by the virus; there is no 
evidence that it plays any role either as cell receptor or in virus pene- 
tration. On the contrary, the inhibitor is found free in the exudate 
from lung and may be partly responsible for the latency of the virus, 
since it may limit virus propagation to the transmission from mother 
to daughter cell and may interfere with its extracellular transfer from 
infected cell to healthy cell (658). 

It should be pointed out that the dissociation of the pneumonia virus 
of mice from tissue inhibitor at low electrolyte concentration is paral- 
leled by the nonenzymatic dissociation of this virus from red blood cells 
at similarly low salt concentrations (658). Influenza virus adsorption 
on red blood cells can also be prevented, and already adsorbed virus 
can be removed, by dilution of electrolytes. These phenomena recall 
the observations on the role of cation concentrations on phage adsorp- 
tion by bacteria (250; 535). 

In summary, Hirst's discovery of hemagglutination by certain viruses 
has led not only to a remarkable simplification of research methods 

CH. 13 Growth Cycle of Influenza Viruses 261 

with these viruses but also to a certain degree of insight into the inter- 
action between virus and cell surface. It has led to the recognition 
of an enzymatic activity in a group of viruses, to the tentative identi- 
fication of their substrates, and to the recognition of the role of these 
substrates in virus adsorption and of their possible involvement in later 
stages of cell invasion. Moreover, it has suggested some leads to the 
therapy of several viral infections, either employing enzymes similar 
to those of the viruses, or using analogues of the substrates for the 
virus enzyme. How far this approach will lead remains to be seen; 
its fundamental importance cannot be doubted. 


The availability of a rapid method for the detection of influenza 
viruses, provided by hemagglutination, has made possible detailed 
studies of the reproductive cycles of these viruses, especially in the 
allantoic sac (299; 305; 353). The situation is somewhat analogous to 
that of a phage growth experiment, with the difference that with the 
animal viruses the susceptible cells are fixed on the walls of the 
allantoic sac. 

Titration of infectivity in the allantoic fluid at various intervals after 
inoculation reveals at first a reduction in free virus ("adsorption"). 
The adsorption proceeds at a constant rate up to amounts of virus 
corresponding to at least 30 infectious units per host cell (122). There 
follows a "latent period" of several hours, during which no increase in 
free virus occurs. Then the virus titer suddenly rises in the allantoic 
fluid. The interval between infection and rise is rather precise, being 
about 5 to 6 hours for strains of influenza A and 8 to 10 hours for 
small inocula of influenza B (304). Heavier inocula of influenza B 
strains give a shorter latent period, approaching that of influenza A 
(427). The latent period decreases with increasing temperatures, be- 
tween 30 and 39 C (659). The rise continues more or less irregu- 
larly until the maximum titer is reached, probably as a result of suc- 
cessive cycles of adsorption and infection of new cells. 

One-step growth experiments. To isolate one cycle of growth, 
Henle and his coworkers (305) utilized the ability of heterologous 
virus (influenza A versus B, and vice versa), after inactivation by 
ultraviolet light, to suppress reproduction of active virus and to reduce 
materially its adsorption onto susceptible cells. Under the proper 
conditions, the heterologous virus does not suppress the growth of 
active virus that has already been adsorbed onto the cells. 

262 Hemagglutination Phenomena 

The experiments are carried out as follows. A first inoculation with 
active virus (for example, influenza A) is followed after 1 or 2 hours 
by a second input consisting of a large excess of UV-irradiated heter- 


3-U 1 1 1 1 I 1 | 1 I 1 I 

1 | 1 | 1 S 1 | 1 

* !l A 

Active A 
PR8 virus 

4 ' " Irradiated Jf + * 
Lee virus 4. / A 

Influenza A strain PR8 

9 } / w 

|3-0- . A ^ / 

j . . * . y 

9 ^ u 5 * 

1.0 - 

1 1 1 1 1 1 i 1 1 I 

i 1 t I i ]( i 1 i 

2 4 68 10 
Hours after 

12 14 i} 24 



Z 3.0 



I ' I 

Lee virus 

PR8 virus 




Influenza B, strain Lee 

i 1 



2 4 6 8 10 12 
Hours after infection 

Figure 86. One-step growth experiments with influenza viruses A and B in the 
allantoic cavity of chick embryos. Heterologous irradiated virus in excess was 
inoculated 1 hour after the active virus to reduce readsorption of liberated virus. 
The ordinate gives the log amount of virus in the allantoic fluid in ID 50 per ml. 
(ID 50 = 50% infectious dose.) From: Henle et al. (305). 

ologous virus (influenza B). Samples are taken at various times, 
antiserum against virus B is added, and virus A is titrated. The virus 
titer rises after the normal constant period (5-6 hours), and a maxi- 
mum titer is soon reached (figure 86). The rise in virus titer in one 
such step of growth is about 60- to 80-fold for strains of influenza A 
and 30- to 40-fold for strains of influenza B. 

A somewhat different picture is observed in experiments in which 
virus readsorption is prevented by the use of RDE instead of heterolo- 

CH. 13 Immature Virus Elements 263 

gous virus and the liberation of virus is followed by taking successive 
samples of fluid from the same eggs (122). Virus liberation, when 
started, apparently continues for many hours; the cells continue to 
excrete virus. The stepwise liberation shown in figure 86 is due to the 
particular conditions of the experiments. The total yield may be over 
1000 per cell. 

The process just described resembles in part the growth cycle of 
virulent bacteriophages, except for the fact that please of virus is a 
continuous rather tb an a snHHpn process^ Further similarities to phage 
growth are actually observed. If allantoic membranes are washed and 
ground early after inoculation, only 1 to 2% of the initial virus activity 
is recovered. There is therefore a definite "eclipse" of the infecting 
virus (299). Most of the residual virus activity is probably due to 
superficially adsorbed virus since it can be destroyed by the addition 
of antibody in the allantoic sac. 

Active virus is found in allantoic membranes collected and ground 
late in the latent period. It begins to appear about 1 hour before 
liberation into the allantoic fluid takes place. 

Interesting results are obtained by following not only the infectious 
virus but also other materials with viral specificity. This is done by 
measuring, besides the infectious titer, the complement-fixing titer and 
the hemagglutinin titer of the allantoic fluid and of extracts of ground 
allantoic membranes. Complement fixation detects both the type- 
specific soluble antigen, consisting of small particles (sedimentation 
constant around 30 S), and other, strain-specific antigens, which are 
found within the virus particles. 

In an allantoic sac infected with influenza A virus, the titer of soluble 
antigen begins to increase in membrane extracts about 3 hours and in 
the allantoic fluid about 4 to 5 hours after infection (figure 87). In 
membrane extracts the rise in soluble antigen may be 1000-fold (353). 
Some time later, the , hemagglutinin titer also rises, both in the mem- 
brane and in the fluid. The demonstration of the hemagglutinin in 
membrane extracts requires removal by RDE of a powerful inhibitor 
of hemagglutination present in the membrane ( 426 ) . 

Thus the complement-fixing antigen and the hemagglutinin appear 
at a time when the infectious titer is still very low. These noninfec- 
tious, virus-specific materials increase in amount before the infectious 
virus. What are the properties of the noninfectious materials and 
their relation to the mature, infectious virus? 

Noninfectious, immature virus elements. Free, soluble antigen is 
present in relatively large amounts in all virus preparations. More of 
it can be extracted from the particles. Noninfectious hemagglutinin 

264 Hemagglutination Phenomena 

has been found to accompany influenza virus and Newcastle virus in 
crude preparations (242; 267). This hemagglutinin is adsorbed by red 
cells and can be eluted. It is not proved that the elution is due to the 
activity of the noninfectious particles themselves, since they have not 
yet been obtained free of accompanying active virus. A fraction of 
the noninfectious hemagglutinin can be separated from the virus by 

| 4.0 

Allantoic Fluid 

Allantoic Membrane 





6 8012 

Hours after infection 

Figure 87. The amounts of various virus-specific materials in allantoic fluids 
and in ground allantoic membranes of chick embryos inoculated with influenza 
virus A. ID 50 = 50% infectious dose titer. Hem = hemagglutinin titer. CFV = 
complement-fixation titer, viral antigen. CFV cone. = same, after eight-fold 
concentration. CFS = complement-fixation titer, soluble antigen. Initial inocu- 
lum 10- ID 50 . From: Henle and Henle (302). 

ultracentrifugation, its sedimentation constant being somewhat lower 
than that of the virus. The hemagglutinin particles contain strain- 
specific, complement-fixing antigen, as do infectious particles. Alto- 
gether, the noninfectious hemagglutinin appears to consist of particles 
very similar to the mature virus particles, but lacking infectivity. 

The hypothesis that the soluble antigen and hemagglutinin, which 
precede the infectious virus in infected cells, are immature stages of 
virus synthesis is supported by other observations. In some cases, 
incomplete virus elements are the only or the main product of infec- 
tion. Massive transfers of high-titer influenza virus into the allantoic 
sac result in the production of allantoic fluids with very low infectivity, 

CH. 13 

Immature Virus Elements 


but with hemagglutinin titer and soluble antigen titer similar to those 
of highly infectious fluids (the Von Magnus phenomenon; 660). Most 
of the hemagglutinin is associated with particles not much smaller than 
the virus (sedimentation constant about 500 S instead of 600-700 S) 
(249; 290). Successive transfers of the allantoic fluids containing 
these "incomplete particles," with a minority of active particles, induce 
further cycles of incomplete reproduction ( see table 25 ) . The hemag- 

Table 25. Serial transfers of influenza A virus in undiluted 
allantoic fluid 

Modified from Von Magnus (660) 

Titer of Allantoic Fluid 







Log Ratio 
/Egg-Infectious Units \ 

\Hemagglutinin Units/ 

















































































glutinin titer remains high and the infectious titer low. After several 
passages, the infectivity titer may suddenly rise again for one or two 
passages. Similar results are obtained using, instead of the intact eggs, 
deembryonated eggs, in which the virus grows in the allantoic cells that 
line the shell cavity ( 71 ) . 

Another case of incomplete formation of influenza virus has been 
observed. Upon heavy inoculation of nonneurotropic strains into the 
brain of mice (582), no increase in infectious titer occurs and the 
infectivity of the inoculum is lost. But there is a rise in hemagglutinin 
titer and in complement-fixing antigen. The inhibitor of hemaggluti- 

266 Hemagglutination Phenomena 

nation that is normally present in mouse brain decreases in amount in 
the course of this abortive growth cycle. The hemagglutinin from 
mouse brain is not much smaller than the active virus particle and is 
possibly similar to the noninfectious particles responsible for the Von 
Magnus phenomenon (290). 

It seems reasonable to consider these abnormal growth cycles of 
influenza as processes in which virus synthesis is arrested at a stage 
where a hemagglutinating but uninfectious particle has been formed 
but not yet been transformed into infectious virus. The incomplete 
growth in the brain suggests that the cells of this organ lack only the 
ability to carry out one specific step needed for virus maturation. The 
potential affinity of influenza virus for the brain tissue is shown by 
other observations: the existence of neurotropic variants, capable of 
reproducing in infectious form in mouse brain (6330); the neuro- 
tropism of influenza virus in the very early chick embryo (289); and 
finally, the ability of influenza virus to interfere with the growth of 
neurotropic viruses in mouse brain (656'; see chapter 14). 

The incomplete reproductive cycle of influenza in mouse brain ac- 
counts for the "toxic" symptoms that follow intracerebral injection of 
virus (298). Damage to brain cells is apparently produced by virus 
that fails to reproduce beyond the incomplete stage. 

The mechanism of influenza virus growth. Hoyle has summarized 
experiments on the production of the various specific materials of a 
strain of influenza A in the allantoic sac and has proposed an elaborate 
theory of virus reproduction (354). According to this author, the two 
types of complement-fixing antigen of the virus, the strain-specific and 
the type-specific ones, play distinct and characteristic roles. The type- 
specific antigen, identified with the soluble antigen, is claimed to in- 
crease logarithmically inside the host cells between the second and 
the fourth hour after inoculation, and to be delayed in growth by dyes 
acting as metabolic inhibitors. The strain-specific antigen, identified 
with the hemagglutinin, increases later and less than the soluble an- 
tigen. Finally, infectious virus is said to appear in the allantoic fluid 
in amounts larger than are ever recoverable from the allantoic mem- 
brane itself. The interpretation offered by Hoyle is that inside the cell 
the infecting virus disintegrates and that the soluble antigen, repre- 
senting the essential viral nucleoprotein, reproduces as such. The 
agglutinin component increases later in the cells, and the two are 
then secreted into the allantoic fluid. In passing through the cell 
membrane, they would produce more agglutinin and acquire a lipoid 

CH. 13 Genetic Recombination 267 

component derived from the cell, binding the other components into 
infectious particles. 

As arguments in favor of this theory, Hoyle mentions first, that ether 
treatment of virus particles frees both soluble antigen and hemag- 
glutinin (presumably by removal of the lipoid layer); second, that 
filamentous forms of virus, when present, are supposedly seen in the 
allantoic fluid but not in membrane extracts, which suggests that they 
may be formed upon liberation; and third, that infected allantoic cells 
removed from the allantoic sac liberate filaments resembling virus 
materials. This is observed both in the dark field and in the electron 
microscope (201; 224). 

This theory involves many assumptions, the justification for which 
is rather weak (246; 426). The acquisition of infectivity by the virus 
upon extrusion from the cells appears plausible, however. The infec- 
tious virus may not be released immediately in the allantoic fluid. 
What we can state with some certainty is that the influenza virus 
particle, upon infecting the host cell, is transformed into noninfectious 
materials. These materials reproduce inside the host cells in the form 
of elements not recognizable as viral particles. The new virus particles 
( or filaments ) are formed by a stagewise process and are preceded by 
different elements (hemagglutinin, complement-fixing antigens) which 
ultimately become incorporated into the infectious virus. For an 
understanding of the relations among the various elements and of their 
relative role in virus growth, chemical studies of various virus-related 
materials are greatly desirable. 

Genetic recombination. Burnet and his coworkers (118a; 120) have 
reported experiments that suggest that influenza virus strains may 
undergo, in mixed infection, phenomena of genetic recombination of 
the type reported for bacteriophage (314). The strains utilized be- 
longed to the influenza A group. Both laboratory variants and natu- 
rally occurring strains were used. Two sets of experiments were first 
reported. In one set, a neurotropic variant NWS, characterized by 
lack of enzyme activity on the egg-white inhibitor, was mixed with a 
serologically different nonneurotropic variant (WSM), characterized 
by heat-resistant hemagglutinin. The mixture was inoculated intra- 
cerebrally into mice. The encephalitic syndrome produced by NWS 
was more or less completely suppressed by interference on the part of 
WSM (27; see chapter 14). From the brains of mice with encephalitic 
symptoms, strains of two new types were obtained. One type, N, is 
neurotropic but more heat resistant and enzymatically more active than 

268 Hemagglutination Phenomena 

NWS. The other type, NM, is somewhat neurotropic, heat resistant, 
and serologically similar to WSM. 

The second group of experiments employed mixtures of strain NWS 
with either MEL or SW, nonneurotropic strains serologically distinct 
from NWS. From the brains of mice inoculated with the mixtures new 
strains were isolated which were neurotropic and resembled, serologi- 
cally and in other properties, the MEL and SW strains. Infection with 
a mixture of NWS and Ocl, a nonneurotropic filament-forming strain, 
gave rise to neurotropic, nonfilamentous derivatives serologically simi- 
lar to Ocl. 

In these and other experiments (118a) the virus isolated from mixed- 
infected animals showed hereditary characters present in the two 
strains used as inoculum. By analogy with the phenomena of genetic 
recombination in bacteriophage, Burnet suggested that genetic re- 
combination takes place with influenza virus in mixed-infected cells. 
The investigators carefully sought to avoid a possible role of selection 
and of strain mixtures in these experiments. They used strains that 
had been made genetically as homogeneous as possible by repeated 
transfers at limiting dilutions, so that each strain had presumably origi- 
nated from a single particle. 

In experiments of this type, in which mass inoculation of animal 
brains leads to repeated cycles of virus reproduction in different cells, 
selection of spontaneous variants cannot be excluded as easily as with 
bacteriophage, where single cycles of infection can be analyzed. More- 
over, the method of transfers at limiting dilutions is not completely 
reliable in eliminating virus mixtures, especially when strains of dif- 
ferent pathogenicity are present. More recent experiments carried 
out in the chick embryo have confirmed and extended the results ob- 
tained in the mouse's brain (120b). 

Reactivation phenomena. Henle has reported (303) that ultra- 
violet-irradiated influenza virus A or B, upon inoculation into the 
allantoic sac, gives rise to a growth cycle whose characteristics depend 
on the amount of inoculum. This suggests, according to Henle, a 
"multiplicity reactivation" similar to that observed in bacteriophage 
(436). Light inocula of the irradiated virus give a growth cycle simi- 
lar to the one that would be expected from an inoculum of active virus 
corresponding to the residual active titer in the irradiated samples. 
With increasing inocula, however, the amount of virus growth increases 
more than proportionally to the inoculum, as though some of the in- 
active virus had been reactivated. It has not yet been shown that this 
reactivation occurs in cells infected with several inactive particles. 

en. 13 Considerations on Virus Reproduction 269 


There are many similarities between the growth cycle of lytic bac- 
teriophages ( chapter 8 ) and that of influenza viruses. We may legiti- 
mately ask whether there is a similar pattern in the growth cycles of 
other viruses as well. 

Certain aspects of virus-host interaction are the same for many 
viruses. Pneumonia virus in mouse lung (254), mumps virus (253), 
and meningopneumonitis virus in the chick embryo (603) exhibit 
growth cycles characterized by an initial reduction (eclipse) of the 
infectious titer below the level of the inoculum, followed by a period 
in which virus titer remains low. The virus titer then increases and 
reaches a maximum after a time interval fairly characteristic for each 

Similar results have been reported for neurotropic viruses (576a). 
An interesting feature is presented by Theiler's virus GD-VII of mouse 
encephalitis. Following peripheral injection in mice, the virus is de- 
tected in infectious form as it proceeds along the nerve fibers toward 
the central nervous system. As the virus reaches the nerve cells, infec- 
tivity disappears for 20 to 30 hours. Later it reappears and reaches a 
maximum around 100 hours after infection, at the time of the onset of 
symptoms. Virus hemagglutinin titer parallels the virus infectivity. 
Here, apparently, the virus loses its infectivity only when it reaches 
the cell body proper, not in the nerve fiber, although the nerve fiber 
is truly a part of the host cell. 

Vaccinia virus, both in rabbit skin and in tissue cultures of the same 
tissue, exhibits a growth curve characterized by eclipse, followed by 
rapid growth (101 a; 158; figure 88). The measurements of infectivity 
are closely correlated with the counts of elementary bodies in infected 

Electron-microscope and light-microscope observations on virus-in- 
fected cells (chapter 11) strongly suggest that the new virus particles 
represent the culmination of a process of stagewise maturation. These 
observations reveal no sign of binary fission of the virus particles as 

The main similarity among all viruses whose growth cycle has been 
analyzed is the eclipse phenomenon, followed by reappearance of virus. 
With some phages, the eclipse has been explained by changes in virus 
organization, especially by the separation of a nucleic acid-contain- 
ing core from a protein skin. With other viruses, the mechanism of 

270 Hemagglutination Phenomena 

eclipse is as yet unknown. With plant viruses, the limited data in the 
literature do not prove the occurrence of an eclipse stage. 

Concerning the later phases of virus development, it should be 
remembered that most information available on animal and plant 
viruses, except that from egg experiments, derives from tests on tissue 
extracts. The cycles studied in this way resemble experiments on pre- 
mature lysis of phage-infected bacteria, rather than cycles of infection 

(b) - 











Figure 88. Growth of vaccinia virus in normal rabbit skin in vivo (a) and in 
shavings of skins in vitro (b). Unfilled circles: no virus recovered. Modified 
from: Crawford and Sanders (158). 

and lysis. Maturation of new virus evidently takes place in the host 
cells, but we know little about the modalities of the virus release and 
of the reinfection of other cells. Symptoms often appear when the 
amount of mature, infectious virus is at a peak; but it is hard to decide 
whether the cell damage reflects the maturation process itself or the 
second wave of attack by newly liberated viruses. Indeed, we know 
that many viruses, and not only the "latent" ones, but some like in- 
fluenza j(326), can reproduce to a maximum extent in susceptible 
tissues without causing any symptom of disease. 

The well-established facts of lysogenicity in phage, together with the 
established similarities between phage infection and infection with 
animal viruses, justify the question as to whether cell infection by 
animal and plant viruses may also be followed by a "lysogenic" type 
of relation, with reproduction of virus in noninfectious form (provirus), 

CH. 13 Considerations on Virus Reproduction 271 

followed rarely ( or even not at all ) by maturation and production of 
active virus particles. 

There are numerous observations that suggest a reproduction of virus 
without maturation. In infections followed by recovery, persistent 
immunity is frequent. This may be due in part to persistence of anti- 
body, but according to many workers this persistence of antibody is 
itself evidence for persistence of virus. In plant viruses, where anti- 
bodies cannot be invoked, infectious virus persists in recovered plants 
and is found in their new leaves. It is present, however, in much 
smaller amounts than in recently diseased plants (see chapter 14). 
One wonders whether the virus does not reproduce mainly in an im- 
mature form, with only occasional maturation. The reproduction in 
immature form might result from the infection of very young tissues, 
just as the frequency of lysogenization in bacteria is affected by the 
physiological conditions of the infected cells. 

The data on latent viruses closely recall the facts of lysogenicity. 
Often, latency simply means absence of symptoms; but there are 
reasons to suspect that sometimes the virus is not present in infec- 
tious form or at least that the infectious virus is only a small proportion 
of the virus present in noninfectious form. Cases like that of herpes 
virus (see 114), which enters the human body in childhood, remains 
unobserved and undetectable (except by antibody test) for many years, 
and occasionally produces symptoms and infectious virus, remind us 
of lysogenic bacteria and of the induction of phage maturation by a 
variety of stimuli. 

It is indeed important to establish whether infection with a virus 
can produce symptoms of cell damage in the absence of virus matura- 
tion. Toxic symptoms of influenza (298) are probably related to in- 
complete maturation. Even in the lung of mice, influenza may cause 
rapid death, with little virus being recoverable in the lung tissue (634), 
an observation that suggests cell damage Without virus maturation. 

Transmission of virus from cell to cell may occur either in the 
noninfectious or in the infectious form. In multi<3^|fcrganisms, 
transmission of noninfectious virus from cell to cell 
occasional maturation, can provide a relation of relative stability;?** Egg 
and sperm transmission of virus diseases, however rare, provides pos- 
sible instances of such a type of relation. 

In summary, we may say that our present knowledge supports the 
idea that virus multiplication inside the host cell proceeds according 
to the following pattern. Upon invasion of the host cell (or of some 

272 Hemagglutination Phenomena 

special part of it) the infectious particle is so modified as to become 
unrecognizable as such. Reproduction of noninfectious virus elements 
takes place by mechanisms, the nature of which still escapes us, but 
which are probably on a level with the reproduction of the specific, 
self -perpetuating elements of the host cell itself (chromosomes, mito- 
chondria, plastids). The role of the preexisting virus in inducing the 
formation of more virus is completely unknown. The cell may possess 
a diffuse, nonspecific machinery for such syntheses, and the specific 
material of the infecting virus may introduce a new model for synthesis. 

Reproduction of noninfectious copies of the infecting virus may be 
followed by maturation, with production of infective particles. In 
bacteriophage, there is good evidence that the matured particles no 
longer reproduce in the cell where they have arisen; that is, before 
reproducing again they must go through the invasion process. The 
mature particle in a sense is a resting stage, which carries the essential 
virus material from cell to cell and probably protects it from external 
injury. For other viruses, the relation of maturation to reproduction 
is not yet clear, but the concept of mature virus as a resting stage seems 
to stand on a solid basis ( 28 ) . Polyhedral viruses of insects reach the 
infectious rod-shaped form by a process of maturation. Their inclusion 
into the polyhedral bodies arrests further production of virus. The 
capsules aroufid the particles of other insect viruses may represent 
protective coats acquired in maturation. 

Reproduction of virus material may lead to a more protracted state 
of immaturity. Even with virulent phages, it is likely that not all 
the phage elements that have been produced reach maturation. It is 
quite conceivable that many viruses, after infecting a host cell, remain 
in that cell (and in its offspring) in a noninfectious form, reproducing 
further or sometimes being eliminated. Maturation may take place in 
some cells only and even in a small minority of the cells. Continued 
reproduction without maturation and without cell destruction may lead 
to cases of nonrecognizable viruses. These will be, both semantically 
and operationally, indistinguishable from the so-called normal cell com- 
ponents; and the problem arises as to the possible origin of certain cell 
components from elements that may have been viruses. The possible 
function of such virus remnants, for example, in normal and abnormal 
development, will be discussed in chapters 17 and 18. 


Interference Phenomena 
in Virus Infections 

In chapters 9 and 13 we have already discussed some of the phenom- 
ena observed when a host is infected with two viruses. Interactions 
in mixed virus infections are of great interest not only for the study of 
virus-host relationships but also in relation to immunity to viruses. In 
bacteriophage it has been possible to analyze interactions between 
viruses in terms of the response of individual host cells. In plant or 
animal viruses this is more difficult, because of intervening complica- 
tions: the multicellular nature of the host, the complicated pathways of 
virus spread in the host, the occurrence of humoral mechanisms of 
defense against viruses. It is a general observation that particles of 
two or more virus types do not react directly with one another outside 
the cells, so that all interactions must have to do with events that take 
place on the surface of, or inside, the host cells. 


Virulent phages. Infection of bacteria with two virulent bacterio- 
phages leads to different results, depending on the degree of relation- 
ship between the two viruses. When the two phages are unrelated 
( according to serological and morphological criteria ) mutual exclusion 
occurs, generally accompanied by a yield-depressing effect (168; 174). 

When the infecting viruses are related, the result depends on the 
time between infections. If two related phages infect the same host 
cells almost simultaneously, they both multiply and mature. The closer 
their genetic relationship, the greater the number of cells which lib- 
erate both phages; for example, all cells mixed infected with a phage 
and one of its mutants, such as T2 and T2r, liberate both types, whereas 
only a fraction of the cells infected with T4 and T6 may liberate both 


274 Interference Phenomena in Virus Infections 

Table 26. Mixed infection with two bacteriophages 

The proportion of bacteria that liberate both phage types following simultaneous 
infection with equal amounts of two phages. 

Phage Pair Per Cent Mixed Yields 














(table 26). In such "mixed-yielder" cells, each phage does not grow 
to the same extent as it would if the other phage were not present; 
rather, the two phages share the total yield, which is approximately 
the same as if only one virus had infected the cell (309). Recombina- 
tion phenomena may take place, leading to production of new virus 
types (173; 314). 

For mixed growth, it is not necessary that two related phages attack 
the host cells simultaneously. The same result is observed if the two 
phage infections are separated by an interval of time during which the 
host cells receive no available nutrient. If, instead, two infections with 
related phages are separated by an interval of a few minutes, while 
the bacteria are actively metabolizing, a mechanism for the exclusion 
of the second phage becomes rapidly established (table 27) (195). 
Indeed, part of the phage that comes late is actively broken down out- 
side the infected cell (238); about 50% of its phosphorus appears in 
soluble form in the medium. 

Temperate phages. A bacterium carrying a phage in lysogenic con- 
dition (prophage) is susceptible to lysis by most unrelated phages 
( although it may be resistant to some individual phage strains ) . Even 
bacteria in which maturation of the prophage has been induced by 
ultraviolet light can still support growth of an unrelated virulent 
phage (669). The virulent phage can actually prevent maturation of 
the prophage. 

Reinfection of a lysogenic bacterium with the phage it already car- 
ries, or with mutants of it, gives rise to complex interactions, including 

CH. 14 

Interference in Plant Viruses 


Table 27. Exclusion between two related phages 

The fraction of bacteria that liberate both T2r+ and T2r phage when infection 
with one phage preceded infection with the other phage by various intervals of 
time. Only about 50% of the bacteria received both phages. Modified from 

Dulbecco (195). 

Time Interval 
Between First 
and Second 

Bacteria That Liberate 

of Mixed 

of Mixed 


r+ only 

r only 

T2r+ first 






1 min 






2 min, 25 sec 


' 428 




5 min, 40 sec 






10 min, 40 sec 






T2r first 






1 min 






2 min, 25 sec 






5 min, 40 sec 






10 min, 40 sec 






prophage substitution, prophage elimination, and double lysogenesis 
(see page 203 and table 23). 

In general, bacteria infected with two phages exhibit two distinct 
categories of phenomena, depending on whether the infection is of the 
lytic or the lysogenic type. In the lysogenic form, a prophage can 
prevent maturation of related phages and limit their establishment to 
the prophage form; the prophage may or may not interfere with un- 
related phages. In the lytic condition, instead, there is no prevention 
of maturation between related phages (provided infection is simul- 
taneous) but there is prevention of maturation between unrelated 


Acquired immunity and cross-protection. In discussing interference 
phenomena among plant viruses, we must consider first what happens 
in plants that are exposed twice to the same virus. In several dis- 
eases, a virus infection, after becoming generalized in a plant, results 
in an apparent recovery. For example, a tobacco plant infected with 

276 Interference Phenomena in Virus Infections 

tobacco ringspot virus, after the early manifestation of ring patterns 
on its leaves (acute stage), seems to recover (chronic stage); the new 
leaves are practically normal in appearance. A second inoculation 
with the same virus does not result, however, in a disease similar to the 
one produced by the primary infection; the recovered plant has an 
acquired immunity toward the virus (see 533). In the so-called re- 
covered plant the virus is still present, although in amounts smaller 
than in actively diseased plants. The ringspot virus in recovered 
tobacco plants is in one-tenth to one-fifth the amounts that can be 
extracted during the active infection. In exceptional cases, it is claimed 
that no active virus can be demonstrated in the recovered, immune 
plants (251). 

Apart from the question whether it is correct to call such plants 
"recovered," and the lack of symptoms upon reinfection "acquired 
immunity," there is no doubt that the previous infection with virus 
prevents the newly introduced virus from reproducing to the same 
extent as it would if inoculated into a virus-free plant. The estab- 
lishment of the immune condition has been attributed to the entry of 
the virus into susceptible cells at a stage of active cell proliferation, 
so that an equilibrium is established between formation of virus and 
formation of cell components. The usual type of infection supposedly 
occurs when the virus penetrates mature cells. 

Situations of this type are quite common with plant viruses, and the 
general observation with plants is that acquired immunity extends not 
only to reinoculation with the identical virus but also to inoculation 
with a closely related virus strain. This cross-protection between two 
related viruses holds also for plants that are in the acute stage of the 
disease. It is particularly clear in the mosaic diseases of tobacco, 
potato, and other plants. The first observations concerned tobacco 
mosaic and its yellow mutants (458; 648b); similar findings were made 
on green and yellow strains of potato virus X. In all these cases, 
previous infection with a less virulent strain, provided it has become 
generalized, prevents manifestations of the symptoms of a more viru- 
lent virus. The opposite type of interference, in which a less virulent 
strain is excluded by previous infection with a more virulent one, is 
more difficult to detect, since the symptoms of the more virulent virus 
do not permit conclusive observations as to the absence of those of the 
milder virus. 

For protection to occur, the cells exposed to the second virus must 
actually be infected by the first. Upon successive inoculation of the 
same leaves with tobacco mosaic virus and with a necrotic variant of 

CH. 14 Interference in Plant Viruses 277 

this virus, necrosis fails to occur in those areas that have been directly 
inoculated with the first virus or into which this virus has already 
spread (figure 89; 386). Here, as well as with severe and mild strains 
of potato virus X, the excluded virus actually fails to multiply in the 
protected area. Protection may sometimes occur upon simultaneous 
inoculation. A high concentration of the ribgrass strain of tobacco 
mosaic virus in a mixed inoculum with ordinary tobacco mosaic virus 

Figure 89. Leaves of Nicotiana sylvestris partly protected against aucuba 
mosaic by direct inoculation with tobacco mosaic. The areas free of the necrotic 
aucuba lesions were rubbed with tobacco mosaic virus 5 days before inoculation 
with the aucuba virus. Courtesy Dr. L. O. Kunkel, Rockefeller Institute, New 

reduces the number of lesions produced by the ordinary strain on bean 
leaves (54a). 

Plant virus studies of this type have been handicapped by the diffi- 
culty of detecting a virus in a mixture with another. This test requires 
indicator hosts susceptible to one virus and resistant to the other, and 
such hosts are seldom available. Moreover, small amounts of plant 
viruses are difficult to detect and titrate. Nevertheless, it seems to be 
well established that in cross-protection the first virus actually inter- 
feres with the multiplication of the second. 

The acquired-immunity phenomena in plant virus diseases might be 
compared with the situation observed with bacteriophages in the lyso- 
genic condition (75). We may suppose that in the recovered plant 
the virus is mainly in a condition (provirus) similar to the prophage. 

278 Interference Phenomena in Virus Infections 

This may explain the nonrecoverability of infectious virus in some in- 
stances (251) and its low concentration in others. Reinfection with 
a related virus would lead to limited reproduction and to failure to 
manifest symptoms. 

Cross-protection and classification. Cross-protection is a satisfactory 
criterion of relationship. Plant pathologists have so concluded be- 
cause they have found cross-protection between virus strains known 
from other evidence to be closely related. Several virus strains have 
been classified by this test (see 533). Although the rule may be 
generally valid, some reservations may be made (460). Cross-protec- 
tion is not always a clear-cut phenomenon. The suppression of the 
second virus may be more or less complete, depending on the amount 
of virus used in the test. A heavier inoculum may simply increase the 
chances of infecting cells not invaded by the first virus, but may also 
allow competitive relations between two viruses in individual cells. 
McKinney has shown that the green tobacco mosaic virus prevents 
manifestation of yellow mosaic symptoms, but that infection with the 
yellow strain does not prevent growth of the green strain, which slowly 
replaces the symptoms caused by the yellow strain with its own ( 459 ) . 
Thus, even with closely related viruses, cross-protection does not 
always work both ways. There is actually one virus that offers a com- 
plete exception to the cross-protection rule, the sugar beet curly top 
virus. This virus exists in several strains differing in virulence; there 
is no protection by the avirulent toward the virulent strains, and even 
an avirulent strain can be shown to multiply in a plant already in- 
fected with a virulent one (250c). 

In some instances, cross-protection may fail to occur between viruses 
that are serologically closely related (46). On the other hand, there 
are rare instances of two completely unrelated viruses (for example, 
severe etch virus and potato virus Y) in which the first virus not only 
prevents infection with the second but also displaces it when it is 
already fully established (44). Similar displacements have been de- 
scribed for diseases, such as those of peaches, that can only be trans- 
mitted by insect vectors or by grafting. Little peach virus displaces 
the virus of peach yellows from yellows-infected buds grafted on plants 
carrying little peach (387). Altogether, the cross-protection tests, 
although generally very valuable, cannot be considered as complete 
evidence for relatedness. 

Virus synergism. It is clear from these few examples that inter- 
ference phenomena in plant virus diseases can take different forms. 
We have described some in which one virus suppresses or inhibits the 

CH. 14 Interference in Plant Viruses 279 

multiplication or the manifestations of another. There are instances 
in which the reverse is true, and two viruses together give more ex- 
tensive and more destructive manifestations than either one alone; for 
example, tobacco mosaic virus gives a severe necrotic disease in plants 
already infected with potato virus X. Sometimes the association of a 
mild virus with a latent virus may produce a rapidly destructive dis- 
ease. Some virus synergism is probably due to an actual increase in 
virus production in doubly infected cells (561a). There might also be 
a more rapid destruction of cells infected by two viruses, or a coopera- 
tion between two viruses infecting different cells in destroying the 
plant tissue. Mixed infection of individual plant cells has been demon- 
strated by the presence of the characteristic inclusions of two viruses, 
one of which gives cytoplasmic, the other intranuclear inclusions (462). 

Acquired immunity and "plant antibodies." A peculiar situation has 
been reported by Wallace for the sugar beet curly top virus in tobacco 
or tomato (662; 663). This virus is transmissible either by graft or 
by the leafhopper Circulifer tenellus. Infected tobacco plants gen- 
erally recover and become immune to reinfection with the same strain, 
whereas tomato plants become severely sick and seldom recover. The 
recovered plants can always be shown to contain virulent virus, be- 
cause a leafhopper can transfer the severe disease from recovered 
plants to healthy tomato. If tomato plants are inoculated by grafting 
from recovered plants, they manifest only a mild disease and recover 
regularly. This situation was interpreted by Wallace as suggesting 
that recovery depends on the production by the plant of a protective 
substance similar to an antibody. In grafting, the protective substance 
is supposedly transferred along with the virus, thus producing a passive 
immunization, whereas the insect transmits only the virus. It is pos- 
sible to imagine, however, that in the recovered plants there exist two 
or more mutant forms of the virus (see chapter 15), some more viru- 
lent, some less, and that the less virulent protect the plant against the 
more virulent. In grafting, the mild virus may have an opportunity 
to spread ahead of the virulent. The insect, instead, may either trans- 
mit only the virulent virus or may transmit both viruses in such a way 
that the virulent one is not hampered by the mild strain. 

Apart from these observations, which, incidentally, could not be con- 
firmed by other workers (533a), there is no evidence for humoral 
mechanisms of immunity in plant virus diseases. Interference phe- 
nomena such as acquired immunity and cross-protection are, therefore, 
especially important, since they seem to be the only mechanisms that 
may lead to a protection from viruses in intrinsically susceptible plant 

280 Interference Phenomena in Virus Infections 

cells. Apparently, most acquired immunity to plant viruses is due to a 
persistent, chronic disease and depends on direct cell protection by 
virus already present within the cells. 


In the study of the responses of an animal host to double infections 
with two viruses or with two inputs of the same virus, we find situations 
more complicated than with plant virus diseases (see 300; 583). The 
major complication derives from the occurrence in animals of serologi- 
cal reactions with the production of antiviral antibodies, whose effect 
on the virus has to be taken into account in interpreting the course of 
infection. Many nonfatal infections with virus (measles, mumps, 
yellow fever) leave a permanent, more or less complete, specific im- 
munity, which may or may not be dependent on persistence of virus 
in the recovered host. Some immunity is transitory, and reinfection 
is possible even after rather short intervals (common cold, influenza). 
This may reflect the rapid replacement of the cells of tissues attacked 
by these viruses ( 556 ) . Usually immunity can be correlated with pres- 
ence of antibodies in the blood serum of the recovered animal. It is 
often difficult to ^establish whether other mechanisms of immunity at 
the cellular level are also involved. 

Acquired tissue immunity. The first point to be considered is the 
fate of the virus during infection and recovery (288). For a number 
x)f viruses, for example, for the infection of the central nervous system 
caused by the herpes virus (466), the amount of infectious virus that 
can be extracted from the infected tissue diminishes after a certain 
time; finally, no active virus can be recovered. This "autosterilization" 
precedes either the death of the host or its recovery, and in case of 
recovery is generally associated with the establishment of immunity 
to a second infection with the same virus. It is not clear whether the 
virus need still be present to stimulate antibody production, thereby 
preventing infection by the second inoculation, or whether a change 
in the antibody-producing cells has occurred, which causes them to 
continue antibody production (or at least to retain the ability to pro- 
duce such antibodies quickly if presented again with the viral antigen). 

A phenomenon that illustrates both the relation between actual 
infection of a group of cells to their immunity upon reinoculation and 
the speed at which such immunity can be established is the so-called 
"rail immunization" (Schienenimmunisierung of the German authors; 
see 287). In herpetic infection of the central nervous system, the virus 

CH. 14 Interference in Animal Viruses 281 

spreads along the nerve fibers from the periphery to the central nervous 
system and, within the latter, from the point of entry to other parts of 
the brain or of the spinal cord. The virus, which if injected directly 
into the brain would cause a rapid fatal infection, often produces only 
a mild infection or even a completely latent one, if it spreads along the 
nerve paths within the nervous system. At the same time, the nerve 
tissue invaded by the virus acquires a definite degree of immunity to 
a challenge inoculation that would otherwise be fatal. It can be 
demonstrated that only those portions of the central nervous system 
that have actually been invaded by the virus are immune to the second 

Phenomena of this kind have been observed not only for neuro- 
tropic viruses but for other viruses as well. For example, influenza 
virus inoculated into mice intraperitoneally soon reaches the lungs, 
where, instead of the usual fatal pneumonia, it causes a mild infection 
and gives rise to solid immunity against reinoculation by the bronchial 
route (553). 

In cases like this, local immunity apparently becomes established as 
the virus spreads, before antibodies are detectable in the general circu- 
lation. One is therefore tempted to suppose that immunity results from 
an alteration of the infected cells, following infection by a route that 
does not lead to their destruction. The nonrecoverability of the virus 
in these immune tissues might have to do either with actual disappear- 
ance of virus, leaving the cells in a modified condition, or with the 
persistence of the virus inside the cells in such a way that it cannot be 
extracted in active form. 

The possibility of the persistence of modified virus inside cells that 
have acquired local immunity might be tested by serological methods. 
For example, we might try to detect virus antigens by the complement- 
fixation reaction in extracts of immune tissues. Such tests have not yet 
been made systematically. 

Attempts have been made to test whether tissues that would be 
immune to a virus in the animal body are still able to support virus 
reproduction if isolated in tissue cultures. The results are somewhat 
contradictory. Some authors report ability of a virus to grow in cul- 
tures of tissues from immune animals, provided antibodies are carefully 
removed (virus III; 24); others report a suppressed or reduced multi- 
plication of viruses in tissues from immune animals (559). It is not 
always clear whether the tissue investigated has acquired local im- 
munity of the type described above, as distinguished from the gen- 
eralized immunity of the animal, which is clue mainly to circulating 

282 Interference Phenomena in Virus Infections 

antibodies. In reduced multiplication of virus in cultures of immune 
tissue, it is also difficult to exclude the presence of some residual anti- 
body in the tissues, since complete removal of antibody is very diffi- 
cult, even by perfusion of immune tissues and organs. 

The theory that local immunity is caused by an alteration of the 
virus-growing ability of the host cells has been weakened by the dem- 
onstration that high concentrations of antibody exist in the very areas 
where virus activity is greatest. In some neurotropic virus infections, 
particularly poliomyelitis and equine encephalomyelitis, the distribu- 
tion of antibodies in the central nervous system closely parallels that 
of the areas of active multiplication of the virus. This parallelism is 
not accounted for by a secondary localization of antibody from the 
circulatory system. The suggestion arises that antibody is produced 
locally by the infected cells; these local antibodies have the same prop- 
erties as antibodies found in serum ( 492; 581 ) . A possible role of local 
antibody production in local immunity (and in rail immunization) is 
supported by the rapidity with which this antibody may be produced 
when mice immunized with the virus of equine encephalomyelitis are 
reinoculated with the same virus (580). The vaccinated mice often 
die if given a small challenge dose, but are fully resistant to intra- 
cerebral inoculation with high virus doses, presumably because the 
high doses cause in the vaccinated animal a quick local production 
of antibodies in amounts sufficient to protect. Immunity may depend, 
in such cases, on the rate of multiplication of the challenge virus; a 
rapidly multiplying strain may develop too fast for the local antibody 
titer to reach an adequate level, whereas a slow-growing strain may be 
held in check. 

These quick local immune responses may explain such paradoxical 
phenomena as the so-called Magrassi protection phenomenon (467). 
Intracorneal inoculation of a neurotropic strain of herpes virus, which 
would normally spread to the brain along the optic nerve and cause a 
fatal encephalitis, fails to do so if the animal receives, 7 or 8 days after 
the first inoculation, a direct intracerebral inoculation of the same virus. 
By itself, the intracerebral inoculation would also cause a rapid, fatal 
disease. Thus, two inoculations, each by itself potentially fatal, mu- 
tually protect the animal when the routes of inoculation and the time 
interval between them are delicately balanced. 

We see then that serological mechanisms may be responsible for 
much local immunity; even where other mechanisms are involved, it is 
difficult to discriminate between their role and that of antibodies. 

CH. 14 Interference in Animal Viruses 283 

Protection by interference. There are instances of acquired resist- 
ance to viruses for which serological mechanisms are not responsible. 
A classic case concerns infection with strains of yellow fever virus. 
A neurotropic strain, restricted in its affinity to the central nervous 
system and almost harmless to monkeys, can protect monkeys against 
the effects of a pantropic strain, which by itself causes a fatal visceral 
infection (221; 350). The neurotropic strain can protect monkeys even 
if inoculated several hours after the pantropic strain. The neurotropic 
yellow fever strain also protects the monkeys against Rift Valley fever 
virus, which is serologically unrelated to yellow fever. 

The above observations illustrate some of the typical features of 
interference between strains of the same virus, and also between sero- 
logically unrelated viruses, where serological protection seems ex- 
cluded. The same is true of interference in the chick embryo, where 
antibody production is absent. 

Interference may be detected by a variety of observations: by sur- 
vival instead of death; by failure of typical lesions to develop; and, 
sometimes, by actual demonstration of a reduced or suppressed virus 
multiplication. Such a suppression can be demonstrated better in 
interference between unrelated viruses, by titrating them separately, 
either in selectively susceptible hosts or after suppressing one of the 
viruses with specific antiserum. Interference between the rapidly 
fatal, serologically unrelated Western and Eastern equine encephalo- 
myelitis viruses can be demonstrated in animals previously immunized 
against one virus; injection of a large amount of this virus gives a 
resistance of short duration against the other virus, due to interference 
(585). The short duration of protection by interference is typical, in 
contrast with the lasting protection given by serological immunity. 

In interference between strains of the same virus, various methods 
can be used to titrate one virus in the mixtures. For example, neuro- 
tropic influenza virus in a mixture with nonneurotropic influenza strains 
can be titrated in the mouse brain. By this method the suppression of 
the neurotropic strain by the nonneurotropic one in tissue cultures 
could be demonstrated (27). 

It seems certain that interference requires the actual occupation of 
a cell by the interfering virus (583). Thus, a virus inoculated in 
excess can seldom be suppressed, except by a more rapidly developing 
virus. The establishment of interference closely parallels the progress 
of infection with the interfering virus, and it wears off as the infection 

284 Interference Phenomena in Virus Infections 

When the interfering virus does not multiply appreciably in a tissue, 
large amounts of it are required for interference. In general, a virus 
is only able to interfere with other viruses in those tissues for which 
it has some affinity. For example, influenza virus can protect the cen- 
tral nervous tissue of mice against Western equine encephalomyelitis 
(656). Although in the mouse brain the influenza virus does not 
reproduce in infectious form, it can at least undergo a cycle of incom- 
plete reproduction. 

The occurrence of interference between serologically unrelated 
viruses is not the rule. Double inoculations may lead either to double 
infection or to interference; the outcome depends on the nature of the 
viruses, on the mode and amount of inoculation, and on the host. 
Relevant data are summarized in tables 28 and 29, reproduced from 
Henle's excellent review (300). In some cases, interference is recipro- 
cal, in others it has been demonstrated in one direction only. 

Dual infection rather than interference may occur with potentially 
interfering viruses if the amounts of viruses inoculated are so small 
that reproduction can take place in different cells (635). In other 
situations, interference fails entirely to take place. With mixtures of 
two viruses, one of which gives intranuclear inclusion, the other intra- 
cytoplasmic, c^lls with dual inclusions have been observed (14; 639). 
Virus synergism, in which a dual infection leads to more severe lesions 
than either infection by itself, has also been reported (219). 

Interference by inactive virus. Interference among animal viruses 
has been observed between inactivated and active virus of the same 
strain, and also between an inactive virus and an unrelated active virus. 
It is a fairly safe prediction that, given the proper conditions, prac- 
tically any interfering ability that a virus exhibits when active will also 
be exhibited by the same virus inactivated by agents such as ultra- 
violet light. Examples are given in table 30. In influenza viruses, 
interference has been observed with heat-inactivated virus that had 
lost its receptor-destroying ability; this indicates that the latter prop- 
erty is not necessary for interference to take place (358). Also, in 
interference between influenza and Western equine encephalomyelitis 
virus, the enzymatic activity of influenza virus is not needed for inter- 
ference (584). 

Large inocula of inactive virus are needed for interference to occur. 
The presumption is that enough virus must be present in the inoculum 
to invade most of the cells in which interference has to be produced. 
Thus, in the allantoic sac, which contains about 10 8 cells, we need at 
least 10 8 -10 irradiated influenza virus particles to suppress the growth 

CH. 14 

Interference in Animal Viruses 


Table 28. Reports on interference between antigenically unrelated 

active agents 

From Henle's review (300), which should be consulted for original references 

Interfering Agent 

Excluded Viruses 

Columbia SK and MM 

Equine encephalomyelitis, 

Equine encephalomyelitis, 


Foot-and-mouth disease 
Herpes simplex 
Influenza A 

Influenza B 
Swine influenza 

Louping ill 

Lymphocytic choriomeningitis 


Newcastle disease 


Rabbit papilloma 
Sheep dermatitis 
St. Louis encephalitis 
Theiler's encephalomyelitis 

Virus III 
West Nile 
Yellow fever 

Poliomyelitis; Western equine encephalomye- 
(?) Newcastle disease 

Eastern equine encephalomyelitis; vesicular 
stomatitis; (?) Newcastle disease 

Rabies; lymphogranuloma venereum 

Rabies; (?) vaccinia 

Influenza B; Eastern and Western equine en- 
cephalomyelitis; Newcastle disease; St. 
Louis encephalitis; Bwamba 

Influenza A; swine influenza; Western equine 

Influenza B; Eastern equine encephalomyeli- 
tis; Newcastle disease virus 


Poliomyelitis; Columbia MM 

Western equine encephalomyelitis 

Western equine encephalomyelitis; (?) polio- 
myelitis; influenza A 

Columbia MM, lymphocytic choriomeningi- 
tis; (?) heterotypic poliomyelitis 

Sheep dermatitis; herpes simplex 

Rabbit papilloma 

Western equine encephalomyelitis 

Western equine encephalomyelitis; rabies; 
poliomyelitis; St. Louis encephalitis; loup- 
ing ill; lymphocytic choriomeningitis; 
lymphogranuloma venereum 

Foot-and-mouth; rabies 

Infectious fibroma 

Venezuelan equine encephalitis; influenza A 

Rift Valley fever; West Nile; influenza A; 
Venezuelan equine encephalitis; dengue 

(?) denotes questionable instances of interference. 


Interference Phenomena in Virus Infections 

Table 29. Reports of dual infection of a host organism or of failure 

of interference 

From Henle's review (300), which should be consulted for original references 

First Agent 

Second Agent 

Eastern equine encephalomyelitis Mumps 

Feline pneumonitis 
Fowl pox 

Herpes simplex 
Infectious myxomatosis 
Influenza A 

Influenza B 


Louping ill 

Lymphocytic choriomeningitis 

Lymphogranuloma venereum 


Newcastle disease 
Pneumonia of mice 
Rabbit papilloma 


Epidemic typhus 

Virus B 
Virus III 
Yellow fever 

Fowl pox 

Herpes simplex; laryngotracheitis; vaccinia; 
feline pneumonitis 

Vaccinia; fowl pox; rabies 

Rabbit papilloma 

Influenza B; mumps; lymphogranuloma ve- 
nereum; epidemic typhus; yellow fever; 
rabies; Semliki forest; pneumonia of mice; 
West Nile 

Influenza A; mumps; pneumonia of mice 

Fowl pox 

Lymphogranuloma venereum 


Influenza A; rabies; louping ill 

Influenza A; influenza B; Eastern equine 
encephalomyelkis; Newcastle disease 


Influenza A; influenza B 

Vaccinia; infectious myxomatosis; virus B; 
virus III 

Heterotypic poliomyelitis; rabies 

Lymphogranuloma venereum; poliomyeli- 
tis; herpes simplex 

Influenza A 

Herpes simplex; fowl pox; virus B; rabbit 

Vaccinia; rabbit papilloma 

Rabbit papilloma 

Vaccinia; Venezuelan equine encephalomye- 
litis; West Nile 

of a simultaneous or subsequent inoculum of active virus (300). It 
has been calculated that one particle can prevent multiplication of 
heterologous virus in a cell (122). 

Interference phenomena of this kind must be taken into account in 
practical problems, as in testing cross-protection after vaccination or 
in assaying inactivated virus preparations for residual active virus. It 

CH. 14 Interference in Animal Viruses 287 

Table 30. Interference between inactivated and active agents 

From Henle's review (300), which should be consulted for original references 

Interfering Agent Excluded Agent 

(inactivated) (active) 

Ectromelia Ectromelia 

Infectious bronchitis of chickens Infectious bronchitis of chickens 

Influenza A Epidemic keratoconjunctivitis; influenza A; 

influenza B; mumps; swine influenza; 
Western equine encephalomyelitis 

Influenza B Influenza A; influenza B; mumps; swine in- 


Mumps Influenza A; influenza B; mumps 

Columbia SK and MM Poliomyelitis 

Newcastle Newcastle 

Psittacosis Meningopneumonitis 

Swine influenza Influenza A; influenza B; swine influenza 

is clear that the presence of an excess of inactive virus in the test may 
mask a considerable amount of residual infectious virus. 

Mechanism of interference. Even though the cellular basis of inter- 
ference among animal viruses is well established, it is seldom possible 
to interpret the observations in terms of phenomena occurring at the 
cellular level, such as mutual exclusion, depressor effects, and preven- 
tion of maturation, which are recognizable in work with bacteriophage. 
Some progress has been made, however, in the study of interference 
between influenza viruses in the allantoic sac of the chick embryo. 
Simultaneous inoculation with influenza A and B, with one virus in 
large excess, leads to suppression of the minority virus; suppression 
also occurs between ultraviolet-inactivated virus and active virus of 
the same type or of the other type. Interference may in part be due 
to enzymatic destruction of the cell receptors for virus adsorption. But 
interference at the level of intracellular reproduction also plays a major 
role; in fact, a second inoculum of inactive virus, whether homologous 
or heterologous, several hours after the first inoculum can still reduce 
the yield of active virus by the infected cells. The second inoculum 
may prevent maturation of those virus particles whose development 
had not reached a certain critical stage (426). 

It is important to realize that interference is observed only when the 
amounts of virus inoculated are adequate. For example, influenza 

288 Interference Phenomena in Virus Infections 

viruses A and B can be carried simultaneously in the allantoic sac for 
at least 9 serial transfers when the inocula are not great enough to 
cause initial interference (635). 

The phenomenon described by Von Magnus (659), consisting of 
the production of hemagglutinating but noninfectious virus upon 
inoculation of influenza virus at high concentration in the allantoic sac, 
may also involve interference. The original virus suspension may con- 
tain a large proportion of immature particles, incapable of reproduc- 
tion but capable of attacking the host cells and of interfering with the 
maturation of normal virus. Upon heavy inoculation, the interfering, 
nanreproducing virus elements may prevent the reproduction of most 
normal virus beyond the stage of the immature particle itself; the cycle 
can then be repeated. 

The interpretation of interference phenomena in animal viruses is 
not yet clear. Apart from cases in which interference may be due to 
removal of virus receptors from host cells, one may conceive of a com- 
petition for substrates, or of a blockade of the processes leading to virus 
production at any one of its hypothetical stages. Usually, if inter- 
ference arrested the development of a virus at the level of the produc- 
tion of intermediate forms, these would not be detected at all. 

In view of the increasing support for the idea that reproduction of 
most viruses involves both a process of reproduction and one of matu- 
ration, it is conceivable that either process may be interfered with at 
any one of its stages. Interference between related viruses may involve 
mechanisms different from those of interference between unrelated 
viruses, just as in bacteriophages. 

The practical significance of interference phenomena. It is difficult 
to assess the role of interference phenomena in the immunity of animals 
to virus diseases. Experimental observations indicate that interference 
phenomena could provide a protection different from the serological 
one. It is debatable, however, whether this cellular mechanism of 
protection plays any important role in the course of natural infections. 

The interference phenomenon has been utilized in the therapy of 
distemper in silver foxes. A ferret-adapted strain of distemper virus 
protects the animal by interfering with the fox-adapted strain (272). 
Protection still occurs when the ferret strain is introduced after the 
virulent strain, provided the virulent strain had been administered by 
the nasal route, which is presumably the natural way of transmission. 

Interference phenomena could be of great significance in virus im- 
munity if viruses remained present in the cells of recovered animals for 
a long time and thereby protected them from reinfection. On the one 

CH. 14 Interference in Animal Viruses 289 

hand, since the best evidence, if any, for the long persistence of viruses 
is based on the lasting presence of antibodies against viruses, it is 
difficult to discriminate between the role of the antibodies and the role 
of the virus itself in protecting against related viruses. On the other 
hand, there is no evidence for lasting immunity against a virus disease 
resulting from previous infections with unrelated viruses. 



Variation in Viruses 

Host Variation and 
Susceptibility to Viruses 


Viruses form a very heterogeneous group and must be considered as 
comprising a number of different and distantly related entities. Great 
variation is found even within groups of viruses that are considered as 
related. Variations are even observed within virus populations raised 
under experimental conditions that make it as certain as possible that 
the whole population has originated from one individual virus particle 
within the duration of the experiment. Thus we witness the appear- 
ance of new virus properties just as we witness the appearance of 
variations in any organism, provided populations of sufficient size are 
observed. Because of this, it has recently become an almost general 
custom among virologists to interpret the variation of viruses, and 
therefore also their evolution, along more or less classic genetic lines 
(114; 218). 

In the past microbiologists had often assumed that viruses (and also 
bacteria) exhibited a peculiarly great variability and adaptability. 
Indeed, it was observed that transfer of a virus to a new host, or even 
inoculation by a different route, was often followed by permanently 
inherited changes in the properties of the virus. This adaptability 
suggested to virus workers a peculiar plasticity of virus heredity. 

Adaptability in populations of higher organisms depends on the 
availability of a variety of hereditary traits, including some that will be 
favored under new environmental conditions. Virus plasticity, instead, 
used to be considered as directly dependent on the ability of the virus 
to change its nature according to the environment in which it grew. 
A priori, it may indeed seem logical to assume that a strictly intra- 


CH. 15 Variation in Bacteriophage 291 

cellular parasite may directly derive some of its heredity from the host 
cell that produces it. 

The belief that most virus variation is directly produced by the host 
environment is unjustified; we now know enough about variation in 
viruses to state this with fair confidence. Just as in higher organisms, 
most (but not all) variation in viruses appears to occur through an 
essentially random production of mutations. The range of mutability 
is determined by the intrinsic potentialities of the genetic material of 
the virus proper. The frequency of such apparently "spontaneous" 
mutations ( mutation rate ) is not obviously affected by extrinsic factors, 
although no adequate data are available on the rates of virus mutation 
in different hosts. 

Virus mutants will be selected in environments in which they repro- 
duce faster, or spread faster, or survive longer than the parent types. 
Microorganisms differ somewhat from higher organisms by the very 
large size of their populations, which, by making available large 
numbers of individuals, provides a great variety of mutants. Selection 
among these can bring to the fore, by selective reproduction, the more 
favorable phenotypes. To give an example, a heavy virus inoculum, 
such as is used in the so-called "adaptation procedures" to make a virus 
grow in a new host, may contain a billion or -more virus particles.* Most 
of these may be unable to grow in the new environment, but if one 
particle in hundreds of millions happens to be a mutant capable of 
growing, it may produce a population consisting almost entirely of its 
own offspring. 

The situation will not Always be so simple. Often, exposure to a 
new environment may lead to no multiplication at all (failure to 
adapt). Sometimes, on the contrary, there may be differential multi- 
plication of two or more types of particles, which either were present 
in the original inoculum or which arose by mutations during growth. 
The composition of the final population depends on the relative abili- 
ties of the different types to proliferate, survive, and spread. All these 
various types of situations are encountered and play some role in the 
survival of viruses in nature. 


Plaque-type mutants. Variation in bacteriophage provides us with 
an almost schematic series of examples on the mode of origin and 
possible consequences of variation in viruses. 

292 Variation in Viruses 

The spontaneous origin of phage mutants is clearly seen in mutations 
the results of which do not require any special selective environment 
to manifest themselves. Typically, populations of certain phages ( for 
example, the coli-phages T2, T4, and T6) always contain mutants 
characterized by plaques larger or smaller than those of the normal, 
wild-type strain (309). Phage derived from these variant plaques con- 
tinues to give the modified plaque type. The frequency of such mu- 
tants is often of the order of 1 in a 1000 particles. If one searches for 
mutants in the phage yields from single bacteria infected with wild- 
type phage, one finds that the mutants, when present, are produced in 
small groups or clones, along with many nonmutant particles (438). 
Besides proving the actual production of mutants during growth, the 
analysis of the clonal distribution of the r mutants, characterized by a 
larger plaque on agar, provides evidence for the exponential rate of 
reproduction of phage material (chapter 9). The frequency of spon- 
taneous mutation of phage T2 from the wild type to the r type is about 
1 mutation for every 10 4 duplications. This value, however, represents 
the sum of the frequencies of all mutations that can give rise to the 
r character; and the number of these is high and as yet undetermined. 

In fact, the genetic determination of plaque type in phages of the 
T2 group is quite complex (see figure 63). The difference in the 
plaque types reflects differences in the extent to which lysis of the 
infected bacteria can be inhibited by a second phage infection (181). 
The least inhibited or r type can arise by any one of a large number 
of different mutations, which apparently occur in different discrete 
genetic determinants, since they manifest genetic recombination (see 
chapter 9). Over 20 strains of T2r, when paired, always gave genetic 
recombination in mixed infection (314). This situation suggests a 
great complexity of the genetic apparatus of the bacteriophage, with 
the presence of a large number of independently mutable elements, all 
involved in the determination of the response to lysis inhibition. 

Host-range mutants. The failure to observe a repeated occurrence 
of the same r mutation in phage T2 is probably not due to any basic 
misunderstanding in the interpretation of the experiments on genetic 
recombination. In fact, similar experiments with other types of mu- 
tants of the same bacteriophage yield different results. This is true 
for mutations affecting the host-range property (host-range or h mu- 
tants). Most of the independently isolated h mutants of phage T2 do 
not give genetic recombinants upon mixed infection, and appear there- 
fore to be mutations occurring at the same genetic site. 

CH. 15 Variation in Bacteriophage 293 

These host-range mutants provide a typical example of the role of 
mutation in virus adaptation to new hosts (434). By plating a large 
population of a bacteriophage with a bacterium related to its sensitive 
host, but resistant to that bacteriophage, we often find a few phage 
plaques. From each plaque we can isolate a new strain of bacterio- 
phage, similar in most respects to the original one but possessing the 
permanent ability to attack the resistant bacterium. The frequency of 
host-range mutants may vary from 1 in 1000 to 1 in 10 or more normal 
phage particles. The mutants may differ from their normal ancestors 
in heat sensitivity and in some other properties, besides host range. 

The host-range mutants of phage illustrate another instructive feature 
of the play of mutation and selection in determining the make-up of 
virus populations. If even one host-range mutant particle is present 
when a suspension of phage is exposed to a resistant bacterium culture 
that the mutant can attack, the mutant will be the only particle to 
proliferate and will completely swamp the normal type and quickly 
eliminate it in successive transfers, since the remaining normal particles 
will be diluted away. 

Mutations may affect any of the phage properties that we can ob- 
serve. Thus, phages that require certain ions or organic cofactors for 
adsorption give nonrequiring mutants (170). Proflavine-resistant mu- 
tants of phage strains have been described (227), as well as tempera- 
ture-resistant mutants (10). Mutations can transform a temperate 
phage into a variety of more virulent forms (100; see table 23). 

Phage mutations observed to date have not affected either the mor- 
phology of the phage particles or the serological specificity of phage 
types, so that these two groups of characters, which are well correlated 
between themselves, form the best basis for classification of phage 
groups. Serological variation with quantitative effects on cross-reac- 
tions is observed both among independently isolated phages of a given 
group and among genetic recombinants (8). 

Host-range characters are of little value in classifying phage, except 
in relation to very unrelated bacterial groups. A phage may acquire 
by one mutation the ability to grow on a strain of a different bacterial 
species, or even genus or family, from its original host. Host range, 
therefore, although representing a useful criterion for identification of 
strains, is too variable a character for taxonomic purposes. 

Population analysis of phage variation. With phages, it is possible 
to carry out interesting analyses of the relative fitness of different 
mutant types growing on the same host (309). Infrequent reverse 
mutations occur from mutant types to the wild type. If any one 

294 Variation in Viruses 

mutant is isolated in pure form and carried through a series of sub- 
cultures, the population tends to revert to the wild type, unless the 
conditions continue to favor the mutants. For example, an r mutant 
will revert to wild type in 3 or 4 subcultures; rare wild-type particles, 
originating by mutations, slowly pile up and are favored by a larger 
yield per bacterium. This is a clear instance of the fact that the virus 
type found as the majority type in a standard situation is generally the 
best fitted for reproduction in that situation. The wild type of phage 
T2 is, of course, the type found in stocks that have been transferred 
repeatedly in the standard medium. The rate of mutation from a T2r 
to a T2 is only of the order of 10 ~ 7 per duplication; but although the 
rate of mutation from T2 to T2r is much greater (about 10 ~ 4 ), it is 
T2 that predominates in a mixture at genetic equilibrium, because of 
its reproductive advantage. In turn, genetic equilibrium can be 
reached quickly because of the large size of the bacteriophage popu- 
lations with which we are dealing. A similar advantage is found for 
the wild type over the h mutants in the presence of bacterial cells in 
which both h and wild types can reproduce; the selective mechanism 
in this case is unknown. 

The reversion of most mutant types to wild type upon subculturing 
has made possible an analysis for independence of various mutations 
(310). A strain that has undergone two successive mutations is sub- 
cultured as described above. If reversion to wild type is always pre- 
ceded by reversion to a single-mutant type, the two mutational steps 
are considered independent; if reversion occurs in one step, the two 
mutations were apparently mutually exclusive, as allelic mutations in 
the same gene would be. 

Genetic recombination. The phenomena of genetic recombination 
in bacteriophage, described in chapter 9, exemplify a type of variation 
peculiar to mixed populations of related phages. New types arising 
by recombination are genetically as stable as new types arisen by mu- 
tation. The role of genetic recombination in phage under natural 
conditions is unknown; by analogy with higher organisms, we may 
speculate on a useful role of recombination in increasing the frequency 
of genetic types within interbreeding phage populations, including 
the populations of prophages (if any such populations exist) in lyso- 
genic bacteria. 

Nonhereditary variation in phage. There are several interesting 
cases of phage variation induced by known factors. In all these the 
variation is nonhereditary, in the sense that the removal of the condi- 

CH. 15 Variation in Bacteriophage 295 

tions that provoked the variation causes an immediate return to the 
original type. 

One example of "phenotypic" variation, as distinct from hereditary 
or "genotypic" ones, is the production of modified phage in bacteria 
that have received a mixed infection with phages T2 and one of its 
relatives, T4 or T6 or T2h (500). For example, cells of Escherichia 
coli, strain B, infected with T2 and T4, liberate part of their phage with 
T4 host range ( active on strain B and an indicator strain B/2, inactive 
on a strain B/4); but, after one cycle of growth on B or on B/2, some 
of this phage becomes stable, normal T2. Phenomena of this kind, 
called phenotypic mixing, suggest that some of the phage produced in 
the mixed-infected bacteria comes out with the genetic material of T2, 
but coated with a "skin" resembling that of the other phage that was 
present in the host cells. In the next cycle of growth the new phage 
comes out with a skin formed under its own control, therefore cor- 
responding to its genotype. 

Another type of nonhereditary variation, already mentioned in 
chapter 9, is a host-controlled modification of the growth ability of 
phage (446). For example, phage P2 normally grows in Shigella 
dysenteriac but does not grow in E. coli, strain B. Occasionally, a 
particle of phage P2 succeeds in growing in some exceptional cell of 
E. coli B. The phage that is liberated by B is in a form P2B, which 
remains capable of successive growth cycles in the E. coli strains as 
long as it stays with this host (76). It is still fully active on Shigella, 
though, and a single cycle of growth on Shigella returns it to the primi- 
tive form, which is almost unable to grow on E. coli. These relations 
are illustrated in figure 68 (page 198). 

In another typical case (446), growth of phage T2 on certain mutant 
strains of E. coli (see table 19) results in the production of phage 
particles which are unable to reproduce in most cells of the E. coli 
host but remain able to grow on certain strains of Shigella. One cycle 
of growth on Shigella restores the normal T2 type, which grows both 
on Shigella and on E. coli. 

There is no question of mutation and selection in here; the variant 
character of the phage is directly and strictly determined by the host 
in which it has grown. 

It appears that, by growing in a new host, the phage particles un- 
dergo some modification of their physiologically active material (but 
not of their genetic, hereditary material), which alters their ability to 
initiate reproduction in some hosts. 

296 Variation in Viruses 

There is no clearly adaptive relation in these changes: some new 
hosts modify the phage so that the phage continues to attack those 
hosts; some modify the phage so as to prevent it from multiplying 
further. We have already discussed the role of host-controlled varia- 
tion in the Vi-phage system of Salmonella typhosa (see page 206). 

The main interest of these phenomena is that they reveal a new type 
of plasticity of viruses and a directly governing role of a host on the 
properties of the virus produced within that host. No definite evi- 
dence for or against the occurrence of such host-controlled variation in 
viruses other than phage has been reported. The relative stability of 
the best-known virus variants with modified host range, even after 
return to their original host, suggest that most host adaptation in viruses 
results from genetic, hereditary changes rather than from host-con- 
trolled, phenotypic ones. But the recognition of the phage modifica- 
tions controlled by the host has opened a new field for the interpre- 
tation of a variety of virus transformations whose mechanism and 
meaning still escape us. 

For example, a virus may reproduce extensively in a certain host or 
in a certain tissue, but apparently only to a limited extent in other hosts 
or tissues to which it is highly pathogenic. Reproduction of active 
virus may even fail completely in diseased tissue, as in virus "masking" 
(see chapter 17). Since the usual tests for virus are made by inocu- 
lation of certain hosts, in some of these complex situations the virus 
may actually be present, but so modified by the host cells that it does 
not reveal itself in the test host. It seems possible that such phenomena 
play some role in the course of the transfer of viruses to new hosts and 
of the spread of viruses from one tissue of the organism to another. 


Variation in plant viruses is evidenced by the occurrence in nature 
of a variety of virus strains for almost every major virus type. These 
strains may differ in the symptoms produced, in host range, and in 
many other properties. 

Tobacco mosaic virus strains have been found in great numbers; a 
systematic collection has yielded close to a hundred different types 
( 362 ) . Most of these are found accidentally in nature, and their recent 
common origin must be inferred by similarities in properties (sero- 
logical cross-reactions; physical similarities between the particles ) . But 
some mutants actually appear under the eye of the investigator. Bright 
yellow patches are occasionally observed in plants infected with a green 

CH. 15 Variation in Plant Viruses 297 

mosaic virus (figure 6). From these patches one can isolate variants 
that breed true and give a yellow, more severe type of mosaic (457). 
The occurrence of yellow mutants in different strains of a virus and in 
different viruses (tobacco mosaic virus and cucumber mosaic virus; 
531 ) suggests the presence of common genetic determinants with com- 
mon mutability in these different viruses. Parallel mutants character- 
ized by similarity of symptoms have actually been found to be asso- 
ciated with virus proteins manifesting similar changes in electrophoretic 
properties from the parental types (243). These viruses with similar 
mutational patterns are probably more or less related to one another, 
although they may have become so differentiated that serological cross- 
reactions may have disappeared. The chemical analysis of several 
closely related mutants (380) shows differences in the amounts of a few 
amino acids. The proteins of serologically unrelated viruses differ to 
a much greater extent (see table 9). 

There is no indication that such differences in composition can origi- 
nate directly as a result of growth in the chemically different environ- 
ments of different hosts. No valid evidence for chemical similarity 
between virus protein and host proteins has been obtained. Mutations 
in plant viruses can reasonably be assumed to arise by the same type 
of mechanism that causes mutations in hereditary material of all other 
organisms. The incidence of mutation can be increased by radiation, 
an agent known to cause gene and chromosomal mutation in plants 
and animals ( 264 ) . Most of the procedures used in obtaining variants 
of plant viruses (passage through different hosts; growth at tempera- 
tures above optimum; selection from areas with unusual symptoms or 
from abnormal local lesions) probably involve simply a selection of 
spontaneous mutants. 

Some interesting conclusions can be derived from ecological observa- 
tions on the distribution of plant virus strains in nature. It is generally 
the mild, nondestructive strains that are widely and more or less per- 
manently established in large natural populations of the host. This is 
only reasonable, since survival and successful propagation are de- 
pendent on the ability of the virus to infect as many hosts as possible, 
allowing them to live and propagate, so that the favorable environ- 
ment to which the virus is bound will not be exhausted. For a para- 
site, invasiveness and destructiveness, the two features of virulence, 
operate in opposition; invasiveness makes for survival of the parasite, 
destructiveness for self-destruction. 

Virulent variants of plant viruses, such as the yellow mutants, are 
found only in exceptional patches at the place of their origin and are 

298 Variation in Viruses 

often difficult to maintain without reversion. For all plant viruses there 
is probably at least one host in which the response is systemic and 
nonnecrotic. This systemic response is clearly more suitable for virus 
survival than the necrotic one. The role of latent virus-plant combina- 
tions in the survival of virus in nature is actually well established. 
Many plant viruses are found latent in symptomless hosts. 


The field of variation in animal viruses includes such an enormous 
number of observations that it would be impossible to cover it in 
detail, even by devoting several chapters to it (see 218). Our purpose, 
however, is to point out the major types of variation, their interpreta- 
tion, and their practical importance both for virus biology and for the 
requirements of therapy, prophylaxis, and control of virus diseases. 
The general principles discussed at the beginning of this chapter make 
it possible to interpret practically every known observation according 
to a single pattern. 

Natural variation is reflected in the occurrence of groups of viruses, 
whose relationship to one another is manifested in the morphological 
similarity of their particles, in serological cross-reactivity, and in 
common patterns of variability. A typical example of such a group 
of related viruses is the vaccinia-variola group. Smallpox or variola 
virus has been known to be the cause of human epidemics for at least 
500 years. In man it causes severe smallpox manifestations; a some- 
what different strain causes a milder disease (alastrim). Horses, sheep, 
and cattle are subject to diseases caused by similar viruses; swine-pox 
is apparently caused by an unrelated virus. Vaccinia virus, the strain 
used for human vaccination, is derived from smallpox either by transfer 
through the skin of calves or, more easily, by successive transfers in 
monkey and rabbit. Other animal pox viruses such as horse-pox also 
give vaccinia strains, characterized by low virulence for the animal of 
origin and by frequent failure to produce typical intracellular inclu- 
sions ( Guarnieri bodies ) . 

The viruses that cause pox diseases in birds, although serologically 
unrelated to the group of mammalian pox viruses, form a similar group 
of agents infectious for several domestic and wild birds. Another well- 
defined group is that of the psittacosis-lymphogranuloma viruses, in- 
cluding several mammalian viruses characterized by similar particles 
and similar growth cycles, and by the susceptibility of the infections 
they produce to treatment by various antibacterial drugs ( 543 ) . 

CH. 15 Variation in Animal Viruses 299 

Among groups of naturally occurring related viruses it is interesting 
to observe the variety of degrees of relationship, as indicated by sero- 
logical reactions. Let us take, for example, the fibroma and myxoma 
viruses of rabbits (602). Myxoma virus contains all or almost all the 
antigenic determinants of fibroma, plus some antigens not present in 
fibroma; fibroma virus cannot remove all antibodies from serum against 
myxoma. The relationship between these two viruses is also shown by 
the fact that both give rise to a similar variant, which produces an 
inflammatory type of disease instead of a tumoral syndrome. 

A more complex situation has been recognized in the influenza 
viruses. Three major types are known: swine influenza, influenza A, 
and influenza B, with weak serological cross-reaction between the 
first two and none between influenza B and the others. A fourth group, 
influenza C, is represented by one strain, #1233. Each major type 
occurs in a variety of slightly different serological types; several such 
types are often isolated from a single outbreak of influenza, suggesting 
a rapid variation in the course of relatively few transfers. It seems 
reasonable to suppose that for a virus that is periodically exposed to 
neutralizing antibodies there will be a selective advantage in a type 
of mutability providing for great serological variation, and more 
especially for variants resistant to antibodies that would neutralize the 
parent type. Thus, hereditary plasticity based on serological variabil- 
ity would afford animal viruses greater chances for successful spread. 

Host adaptation and tissue adaptation. Virological literature is full 
of examples of virus variation arising under the eye of the experimenter. 
The two most commonly observed types, because of their importance 
in practical work, are host adaptation and tissue adaptation. A virus 
pathogenic for man will often give no manifestation, or even not 
reproduce at all, if injected into a laboratory animal or into the chick 
embryo; but it may become adapted to do so by repeated transfers, 
especially by the use of large inocula. This type of adaptation is ob- 
tained more or less easily for different viruses and is probably due to 
the selection of variants which are capable of reproduction in the new 
host. Typical examples are the adaptation of rabies virus to rabbits, 
the adaptation of human poliomyelitis virus to mice (something that 
has occurred only exceptionally, for instance, for the so-called Lansing 
strain), and the adaptation of a number of viruses to growth in the 
chick embryo. The adapted variants obtained in this way are gen- 
erally stable and maintain their new host range even after some trans- 
fers through the original host. 

300 Variation in Viruses 

Tissue adaptation follows the same pattern and consists of the 
"training" of a virus to reproduce in a tissue different from the one 
in which it was found. For example, the viruses of vaccinia, yellow 
fever, herpes, influenza A, and Newcastle disease give rise to neuro- 
tropic variants, which can be revealed by injecting large amounts of 
the virus directly into the central nervous system of susceptible animals. 
Examples of host and tissue adaptation could be multiplied at will, a 
major segment of the experimentation on animal viruses being devoted 
to finding more suitable hosts for virus cultivation. 

Changes associated with adaptation. It is important to keep in mind 
that, since adaptation probably consists in the selection of spontaneous 
mutants, the adapted strains may differ from the strain of origin not 
only in the new host or tissue affinities but also in other properties. 
This is only natural, because virus mutations are likely to be pleiotropic, 
that is, to affect a number of different characteristics. For example, 
neurotropic vaccinia is more virulent for the skin and for other organs 
of rabbit than the dermotropic strain (190). Neurotropic strains of 
yellow fever, on the contrary, after long selection by intracerebral 
transfer, are almost completely nonvirulent for most organs of monkey 
and presumably of man. It was a neurotropic strain of yellow fever 
which, after more than 50 transfers in chick embryo tissue cultures, 
gave rise to Theiler's strain 17D, which is not pathogenic for man but 
immunizes against yellow fever. This strain is now used the world 
over for vaccination ( 647 ) . 

Another classic case is that of rabies virus. The virulent virus from 
rabid dogs ("street" virus) was attenuated by Pasteur by means of 
successive transfers through the central nervous system of rabbits. 
This attenuated virus ("fixed" virus), after proper treatments, can be 
used for human vaccination. Innumerable examples can be found in 
any book on medical virology. Indeed, the selection of nonvirulent 
but immunizing virus strains by means of adaptation procedures has 
provided the best vaccines for the prophylaxis of virus diseases. 

A number of interesting changes are associated with the early stages 
of adaptation of influenza virus to new hosts (118). Human strains 
as first isolated from throat washings or after some transfers in the 
ferret lung are in the O form, characterized by a low ratio between 
their hemagglutinating titer for red blood cells of fowl and that for 
guinea pig red blood cells (low F/G ratio). After several transfers 
through the allantoic or the amniotic cavity of the chick embryo, the 
F/G ratio increases. The increase is apparently due to an actual re- 
placement of the O particles by a D mutant, which is better suited for 

CH. 15 Evolution of Virus Diseases 301 

growth in the chick embryo. The virus can be maintained in the O 
form for several transfers in eggs by using inocula with very little virus. 
Presumably, the virus population produced in an early egg transfer 
still consists predominantly of the original type. A light inoculation 
will generally transfer only this type, and the process of variation and 
selection of mutants has to start all over again. If, however, the inocu- 
lum happens to contain some of the variant virus type, then this type 
will suddenly take over. In fact, even small inocula can maintain the 
O form for 5 or 6 transfers only. 

The change from O to D form also results in several other changes, 
affecting pathogenicity, extent of multiplication, and susceptibility to 
inhibitors of hemagglutination. Other variant forms of influenza viruses 
appear upon egg transfer (613). Important changes also take place 
in influenza virus during the first few transfers through mice. Strains 
isolated from the same epidemic outburst in man, after passage through 
mice, are often serologically different. If these differences are already 
present in the original strains from the patients, they have to be taken 
into consideration in the epidemiological interpretation of such out- 
bursts (231). Since, however, some serological variation accompanies 
the adaptation of virus to mice, an originally homogeneous group of 
isolates from a human epidemic may give rise to a serologically hetero- 
geneous group of viruses after mouse adaptation ( 326 ) . 

Virus variation and the evolution of virus diseases. The few 
examples of variability that we have mentioned give us a lead to the 
interpretation of the origin of naturally occurring groups of more or 
less obviously related viruses; each group is probably derived from a 
common ancestor, which might or might not have been identical with 
one of the forms currently encountered. Generally, all viruses of a 
group have morphologically similar particles, in line with the findings 
on bacteriophages. In looking for the common ancestor of a group 
of viruses we should expect it to be a virus that produces a mild or 
fully latent disease of some host, while having opportunities to come 
into contact with the other hosts now parasitized by the group of 
viruses. A good example is that of the mammalian pox viruses (117). 
One member of the group is the virus of ectromelia, a mild endemic 
disease of mice. The ancestral type of the group may have been either 
ectromelia itself or some other form of virus well established in some 
rodent. Rodents are widespread, come in close contact with many 
other mammals, and are almost ideally fitted to act as virus reservoirs. 

The problems of virus evolution and of the survival of viruses be- 
tween epidemics are among the most fascinating aspects of virology 

302 Variation in Viruses 

(114). The record is extremely fragmentary, and evidence is often 
supplemented by guesswork. Even so, a few examples illustrate the 
principles. Yellow fever as a human disease would have been eradi- 
cated by mosquito control had it not been for the occurrence of a 
jungle form of the disease in wild monkeys and other animals in the 
tropical forests. Psittacosis, sporadic and very virulent in man, is 
endemic among parrots and other birds. It is maintained not only in 
wild birds but also in domesticated ones, because of transmission from 
mother to young in the nest. Infection of the young is frequently 
followed by apparent recovery, with return to a stage of active lib- 
eration of virus after the animals reach maturity (483). 

Some instances of variation in animal viruses involve complex situa- 
tions still unclarified. One such is the fibroma-myxoma virus trans- 
formation (73; 612a). Fibroma virus causes a mild disease of rabbits, 
characterized by fibrous tumors, which generally regress; it may be 
the ancestral form of this virus group. The highly virulent myxoma 
virus, characterized by a jellylike transformation of connective tissues, 
is always fatal to rabbits. Injection into the rabbit skin of a mixture 
of live fibroma virus and heat-killed myxoma virus (which cannot 
cause myxoma by itself) often gives rise to typical myxoma. This 
indicates either a transformation of the fibroma into myxoma virus or 
a reactivation of the inactive myxoma. Successful transformation de- 
pends on the relative amounts of live fibroma and heated myxoma 
virus. The transformation can also be induced by deproteinized ex- 
tracts of myxoma (612a). 

We have already discussed in chapter 13 the recent experiments on 
phenomena resembling genetic recombination in influenza viruses 


The host range of any virus must reflect not only the genetic prop- 
erties of the virus, but also those of its hosts. Different tissue affinities 
must reflect differences arising in the course of development among 
the cells of an organism. Usually the genetic and developmental 
factors that determine virus-growing ability are too complicated for 
analysis in either genetic or biochemical terms. Some simple cases of 
differences in virus sensitivity or in virus response among closely related 
organisms offer an opportunity to see some of the basic mechanisms 
in action. 

CH. 15 Factors in Virus Sensitivity 303 

Sensitivity to bacteriophage. With bacteriophages, the sensitivity 
of a bacterial host may be suppressed by a single spontaneous muta- 
tion (435). That the bacterial mutations to phage resistance are spon- 
taneous, that is, not induced by contact with the phage, is shown, for 
example, by the fact that phage-resistant mutants can sometimes be 
isolated without previous contact with the phage, either because of 
peculiarities in colonial morphology or by means of special detection 
procedures (411). 

The spontaneous origin of the bacterial mutations to resistance re- 
sults in a peculiar distribution of mutants in a series of similar bacterial 
cultures (442). The sooner the first mutation to resistance happens 
to occur in a given culture, the larger will be the number of mutants 
in that culture, when later tested. If many similar cultures are tested 
for the proportion of phage-resistant mutants they contain, they show 
wide fluctuations. These fluctuations are, therefore, an expression of 
the fact that the mutants are already present in the cultures, in the 
form of clones of identical sibs, before the cultures have ever come in 
contact with phage. This forms the basis of the so-called "fluctuation 
test" for spontaneous bacterial mutations. 

The genetic resistance of bacteria involves either complete inability 
to adsorb phage or only inability to carry out the irreversible stage of 
phage adsorption (250). It has proved valuable material for the study 
of bacterial genetics (435). A bacterium may by one mutation be- 
come resistant to one or more of the phages that attack it, and, by 
successive independent mutations, may acquire resistance to several 
phages. There seems to be no correlation between the properties of 
two phages with a common host and the existence of bacterial muta- 
tions producing joint resistance to both phages. A mutation to re- 
sistance to one phage often modifies several bacterial properties, such 
as ability to adsorb other phages, ability to synthesize some essential 
metabolites, or ability to produce other phages in normal form. For 
example, certain mutants of Escherichia coli, strain B, are resistant to 
phage T4 and, although still sensitive to phage T2, produce a modified 
form of T2 (446; see page 295). 

Bacteria with known sexual phenomena of the genetic-recombination 
type (specifically, E. coli, strain K-12) show that each mutation pro- 
ducing phage resistance behaves as a change in a single gene. In 
diploid races of K-12 heterozygous for resistance and sensitivity to a 
phage, sensitivity is dominant (410). In this bacterial strain, even the 
alternative characters, lysogenicity or sensitivity for the phage lambda, 
behave as a pair of allelic differences (409). 

304 Variation in Viruses 

Plant viruses. Good evidence has been presented for a simple 
genetic determination of the type and severity of plant response to 
several plant virus diseases. The best examples are those of genetic 
control of necrotic response to tobacco mosaic virus in tobacco and in 
other plants. 

The usual tobacco plant, Nicotiana tabacum, as well as many other 
species of Nicotiana, respond to tobacco mosaic virus by a systemic 
chlorotic disease. Nicotiana glutinosa responds with necrotic lesions 
at the points of entry of the virus. This localization of the virus pre- 
vents its spread. If present in the commercially valuable N. tabacum 
it could confer a "field immunity" to a tobacco crop. Hybrids of 
N. tabacum X N. glutinosa give the necrotic response; this is due to 
a dominant factor (N) from N. glutinosa (337). These hybrids are 
sterile and cannot be used to transfer the N gene to N. tabacum by 
further crosses. Success was obtained (3390) by crossing N. tabacum 
with N. digluta, a fertile artificial amphidiploid hybrid ( carrying both 
the tabacum and the glutinosa chromosomes). Some of the progeny 
plants had the N type of response; unfortunately, this was present in 
a chromosome derived from N. glutinosa (250a) and was accompanied 
by other commercially undesirable glutinosa properties. Further 
crosses have succeeded in reducing the amount of glutinosa heredity 
in the tobacco derivatives and seem to have provided commercially 
valuable derivatives (250b; 654b). The resistance conferred by the 
necrotic response is not complete, since generalized necrosis may follow 
heavy inoculation, but is quite adequate under field conditions. The 
genetic basis of the necrotic type of resistance to tobacco mosaic virus 
in plants other than tobacco has also been elucidated ( 339 ) . 

Another type of resistance has been observed in the variety Am- 
balema of N. tabacum. This variety remains symptomless if inocu- 
lated with tobacco mosaic virus (498a). The dominant factor for 
resistance (A) can be introduced in other varieties of tobacco. Its 
practical usefulness may be limited by the fact that the virus may 
multiply in the resistant plants. 

Single gene differences determine the ability of an insect ( Cicadulina 
mbila) to transmit the corn streak virus; the dominant allele of a sex- 
linked gene determines ability to transmit the virus, apparently by 
controlling the permeability of the wall of the insect gut for the virus 
(632). Other cases of the same type exist (see chapter 16). 

Animal viruses. Genetic work on sensitivity to viruses in higher 
animals has been limited by the practical difficulties of pure-line work 
in mammals; until recently there was an inadequate realization of the 

CH. 15 Factors in Virus Sensitivity 305 

importance of hereditary factors in resistance to disease (see 265). 
"High-mortality" and "low-mortality" strains of mice with respect to 
several viruses (pseudorabies, equine encephalitis) have been reported. 
Some results of hybridization suggest a single-gene control of suscepti- 
bility and resistance; more commonly, control is probably polygenic. 

A discovery by Sabin (572) has provided a clear-cut example of 
genetic determination of sensitivity to animal viruses. An inbred strain 
of mice (PRI) is 100% resistant to the neurotropic 17D strain of yellow 
fever, and is also resistant to several other neurotropic viruses ( dengue, 
West Nile fever, encephalitis of the Japanese B, St. Louis, and Russian 
types; not to poliomyelitis, equine encephalomyelitis, or rabies). In 
crosses with other inbred mice, the resistance of strain PRI behaves as 
if controlled by a single dominant gene. Resistance is only complete 
in the adult mice, not in the newborn. The mechanism of resistance 
is a reduction in the extent of production of infectious virus in animals 
with the dominant gene. It is not known whether virus multiplication 
is altogether suppressed or arrested at some noninfectious stage. 

Other studies of this kind have seldom given such clear-cut results. 
These studies are complicated, in the first place, by intrinsic complexity 
of the genetic mechanisms involved (multiple-gene characters; genes 
with incomplete penetrance, that is, with more or less frequent failure 
to produce their phenotypic effect). In the second place, there are 
complications due to developmental differences. These manifest them- 
selves both in the various tropisms of viruses, which cause them to 
reproduce in some tissues of a host and not in others, and in the 
influence of age on the susceptibility of the organism as a whole. 

Thus, even a bacterium may differ in its response to a phage accord- 
ing to the stage in its cultural cycle, from actively reproducing cells to 
resting cells. In plants, a virus that invades rapidly growing cells may 
produce a milder infection than in mature cells. The competition of 
actively synthesized protoplasmic components may be too strong to 
allow the virus to grow as much as in nongrowing cells. 

In animals, the age of the host may affect either the tissue response 
or the successful operation of immunity mechanisms. Newborn ani- 
mals are inefficient producers of antibodies; in man this condition lasts 
about 6 months after birth and is reflected in greater morbidity and 
mortality, for example, from encephalomyelitis (579). Persistence of 
antibodies from immune mothers may sometimes play a role in the 
opposite direction. In mice, the degree of immunity produced by 
vaccination with Eastern equine encephalomyelitis increases rapidly in 
the first few days after birth. This age-dependent immunity is strictly 

306 Variation in Viruses 

peripheral and due to production of antibodies; the susceptibility of 
the brain to direct inoculation with the virus remains constant (491). 

In some cases, the age of an animal determines its receptivity to a 
virus, which then remains latent for a long time in the animal's tissues. 
This is so, for example, for the so-called milk factor of mammary 
carcinoma in mice (see chapter 17). Age and hormonal factors de- 
termine the later manifestations of the latent virus. 

Viruses and tumor al host cells. An important application of the 
specific tropism of viruses is their use as tumor-destroying agents 
(489a). A variety of viruses will attack preferentially the cells of 
certain tumors and destroy them. For example, a mouse sarcoma may 
become necrotic and nontransplantable 6 days after the mouse that 
carries it is inoculated intraperitoneally with encephalitis virus (Rus- 
sian strain ) . A virus may even destroy a tumor of an animal in which 
it produces no sign of disease, for example, encephalitis virus in tumor- 
bearing chickens (594). 

This "oncolytic" ( = tumor dissolving ) property of viruses obeys 
definite rules of specificity. Certain viruses destroy only certain tumors, 
others none. There is no absolute relation between the ability of a 
virus to destroy a tumor and its ability to reproduce in its cells. Suc- 
cessive transfers of a virus in a tumor may increase its oncolytic ability 
(489a); the mechanism of this adaptation to a more destructive action 
on the host cell has not yet been analyzed. 



Transmission, Vectors, 
and Survival of Viruses 

For its successful propagation, a virus must come in contact with 
susceptible cells. We must now consider how virus propagation is 
effected in nature. A consideration of virus transmission and survival 
furnishes the key to an understanding of the epidemiology of virus 
diseases, hence to a rational prophylaxis. 

Transmission presents no special problems for bacteriophages. 
Chance contact between phage and bacteria is probably the general 
mechanism of propagation. Lysogenic bacteria are the virus reser- 
voirs, and it is likely that any phage well established in nature exists 
in lysogenic relation with some bacterial strain. 

The situation is more complex for animal and plant viruses. The 
external surfaces of the host organisms act as barriers to virus pene- 
tration. The structural differentiation of the host often buries the sus- 
ceptible cells in deep tissues and organs. To parasitize successfully 
an animal or a plant, a virus must penetrate the surface and sometimes 
also internal barriers, reach the susceptible cells, multiply, and be re- 
leased in such a way that it can reach another host. 

Direct mechanical transmission is possible only through wounds or 
by contact with poorly protected cells. In animals, the cells lining 
the respiratory or digestive tract are the most exposed. Often, trans- 
mission by mechanical means is inadequate and the virus must be 
introduced directly beyond the protective cell layers, generally by 
arthropods or other vectors. 

Once penetration has been accomplished, the virus often has a long 
way to go in order to reach susceptible cells. Success is possible be- 
cause of the existence of spreading pathways. Plant viruses move 
along with food in the vascular tissues of plants. Likewise, animal 
viruses circulate with the blood, the lymph, and possibly the spinal 


308 Vectors and Survival of Viruses 

fluid l of animals until they reach susceptible cells. Humoral trans- 
port is especially successful in the early phases of infection, before 
circulating antibodies appear. 

Some viruses, however, have special ways of spreading. Several of 
the neurotropic viruses spread along the nerve fibers to the central 
nervous system (352). This nerve transport appears to occur in the 
axon rather than along the surrounding sheaths of the nerve fiber. 
This is the major route of spread for some viruses, which may reach 
the nerve endings in the mucous membranes or in exposed, wounded 
tissues (93). Not all neurotropic viruses are restricted to the nerve 
pathway, however. Poliomyelitis virus apparently multiplies at the 
point of entry and in other tissues (573), and is found transitorily in 
the blood circulation. Its spread to the central nervous system may 
occur either through the blood or along the nerve fibers (91). Spread 
through the blood before reaching the central nervous system is the 
rule with equine encephalomyelitis. 

Nerve transport occurs both centripetally and centrifugally, as well 
as from one part of the central nervous system to another. It is 
doubtful whether viruses such as rabies and poliomyelitis actually 
multiply in the nerve fibers. The rate of progress of poliomyelitis 
virus in nerves (2.4 mm per hour; 92) is compatible with the idea of 
spread without multiplication. 

The precision and selectivity of the mechanisms for virus spread in 
an infected organism may seem extraordinary. We must, however, re- 
member that the possibility of infection exists only when such mech- 
anisms are present. Virus-host relations as we find them in nature are 
the product of a long process of selection, which continuously perfects 
these relations by trying the viruses in a variety of anatomical and 
physiological environments. 

Mechanical and graft transmission of plant viruses. Few plant 
viruses are regularly transmitted by mechanical inoculation. Tobacco 
mosaic virus can penetrate through the roots, although by this route 
it often fails to invade the rest of the plant (532). Infection of roots 
from soil also occurs for tobacco necrosis and wheat mosaic viruses. 
Some of the potato viruses are perpetuated by the planting of infected 
tubers. Seed transmission is the exception rather than the rule. Even 
when tuber and seed transmission can maintain viruses, transmission 

1 The spinal fluid probably transmits viruses injected into it to the central 
nervous system but does not seem to be involved in the natural propagation of 
neurotropic viruses. 

CH. 16 Mechanical Transmission 309 

from diseased plants to healthy plants is necessary for virus survival; 
otherwise the diseased plants would ultimately be displaced by healthy 
ones and the virus would perish with them. Therefore, all viruses 
must be transmitted in nature from plant to plant by external mecha- 
nisms; most of them are spread by insect vectors. 

Practically all plant viruses are transmissible by grafting (or bud- 
ding), but this method plays a role only in the natural spread of 
viruses of domestic fruit trees. Plant viruses can also be transmitted 
through dodder. 

Mechanical transmission of animal viruses. Many animal viruses 
are regularly transmitted by mechanical transfer. We may mention 
measles, psittacosis, poliomyelitis, the mammalian pox viruses, and 
pseudorabies. These viruses are liberated from infected individuals in 
crusts, droplets, or feces, from which they can easily reach accessible 
mucous membranes of other individuals. For poliomyelitis, the fre- 
quent presence of virus in the feces of infected and carrier persons has 
incriminated flies (478) and cockroaches (356) as possible vectors. 
These insects would be vectors only in the sense of spreading the virus 
and introducing it into food, not as specific carriers involved in the 
actual propagation and inoculation of the virus. The inoculation of 
rabies virus by the bite of rabid animals may be considered as a 
peculiar instance of mechanical transmission. 

Human-disease viruses that are transmitted mechanically will sur- 
vive if they produce mild diseases that are widespread in the human 
population. For example, the virus of lymphogranuloma venereum 
remains in the infected person for a long time, is transmitted almost 
exclusively by sexual contact, and is very frequent among sexually 
promiscuous groups in certain localities. Another survival pattern is 
exemplified by those viruses sporadically causing a serious disease in 
man, while the virus is endemic in a domestic or readily domesticated 
animal. For example, lymphocytic choriomeningitis is common in 
grey mice (31 ), in which it remains active for a long time and causes 
a symptomless infection transmissible from mother to offspring in the 
womb (651). In such cases, even a high mortality rate in an acci- 
dental host, such as man, does not preclude virus survival. The epi- 
zootic pseudorabies virus causes a regularly fatal disease in cattle. It 
persists in nature because of its widespread occurrence in hogs, in 
which it causes a readily transmitted, extremely mild syndrome (599). 

Control of human viruses transmitted mechanically is even more 
difficult than control of insect-transmitted viruses, as evidenced by the 
persistence of measles and poliomyelitis. Mass vaccination is the only 

310 Vectors and Survival of Viruses 

effective measure, provided effective vaccines are available, as for 
smallpox. (In domestic animals, breeding for immunity is an impor- 
tant possibility.) The futility of many quarantine measures (351) is 
due to the mild and often symptomless course of the infection in most 
individuals (subclinical infections). 

Poliomyelitis is again pertinent (114; 351). The paralytic disease 
is apparently the exception, probably representing about 1% of the 
cases of infection; a genetic factor seems to be involved in its occur- 
rence. Most people have contacts with the virus in childhood, pos- 
sibly more than once, and become at least partially immune for life. 
The maximal incidence of primary infections seems to have shifted 
from infancy to early school age. This may be due to a later average 
time of first exposure to virus in children, because of a higher propor- 
tion of immune adults in the community. In spite of this widespread 
immunity, there must always be present enough subclinically infected 
carrier individuals, probably children with primary infections, to give 
the virus its almost world-wide distribution in the form of small epi- 


Many animal viruses and practically all plant viruses are transmitted 
in nature by arthropod vectors, mainly insects. This mechanism of 
transmission is favored by the feeding habits of the vectors. Blood- 
sucking insects such as mosquitoes and lice take up a virus by feeding 
on infected animals and transmit it to other animals of the same spe- 
cies or of other species. In plants, sucking insects (aphids, leaf- 
hoppers, etc. ) and some biting ones are continually feeding on plants 
and moving from one plant to another, thereby transmitting viruses. 
Some such vector mechanism is usually a necessity ( see table 2 ) . 

Animal viruses. As examples of vector transmission we may take, 
first, yellow fever and dengue fever, both of which are transmitted by 
mosquitoes, particularly by the species Aedes aegypti. This is a fre- 
quent pest in tropical and subtropical human communities along the 
coasts of the Atlantic Ocean and the Mediterranean Sea. The proof 
by Walter Reed and his group (548) of the role of mosquitoes in the 
transmission of yellow fever has long been part of the saga of medical 

The equine encephalomyelitis viruses ^re transmitted by several 
mosquitoes and other arthropods. The Western form of equine en- 
cephalomyelitis (WEE) can be transmitted by a common tick, Der- 
macentor andersonii. The Eastern form ( EEE ) and a Russian form 

CH. 16 Transmission by Vectors 311 

are also transmitted by ticks, and so are the louping ill virus of sheep 
and the virus of Colorado fever. Sandfly fever (papataci fever), a 
mild virus disease common around the Mediterranean, is transmitted 
by a diminutive fly, Phlebotomus papatasii. 

The relations among virus, host, and vector are ecologically and bio- 
logically complex. Not all arthropods feeding on an infected host can 
transmit viruses, and different insects transmit different viruses. More- 
over, if we find that a given insect can transmit a virus, this does not 
mean that it is the vector of that virus in nature, or even one of the 
usual vectors. Ecological factors may completely prevent it from 
playing a role. For example, although ticks can transmit the Western 
equine encephalomyelitis virus (638), it is doubtful that they play 
any significant role in its spread. 

After feeding on an infected animal, the vector is not always imme- 
diately capable of infecting another animal. For example, it takes 
8-12 days for Aedes aegypti to become "infectious" after engorgement 
with yellow fever-infected blood (419), and 7-10 days after feeding 
on a dengue patient (see 87). What is the reason for this delay or 
"incubation period*? One reason is that the virus finds itself in the 
gut of the insect and must reach the salivary glands before being in- 
jected with the insect's saliva into another susceptible animal. The 
passage from the gut to the blood and from the blood to the glands 
may take some time. There is definite evidence that the intestinal 
wall of an insect may be a barrier to virus penetration. Aedes aegypti 
can transmit WEE virus, but apparently not EEE virus. Merrill and 
Tenbroeck (481), repeating with the EEE virus a classic experiment 
on a plant virus (see page 315), found that if, after feeding on emul- 
sions of guinea pig brain containing EEE, mosquitoes had their ab- 
domens punctured, so that the ingested virus could penetrate into the 
body cavity, they became capable of transmitting the virus. In viru- 
liferous mosquitoes the virus is present in practically all tissues and 
organs (342). 

Multiplication of animal viruses in vectors. At least some animal 
viruses multiply in the tissues of the vector. The incubation period 
represents in part the time needed for multiplication of the virus. 
Yellow fever virus can be titrated by injecting mice intracerebrally 
with the pulp of Aedes aegypti at various intervals after the insects 
have fed on a monkey infected with yellow fever. The virus titer 
decreases at first, then increases and reaches a level that may be 50 
times higher than the initial one (681). This rise in titer is particu- 
larly evident if the mosquitoes are not allowed to ingest too much 

312 Vectors and Survival of Viruses 

virus (a mosquito can ingest as much as 10 infectious units at one 
feeding). With the same mosquito, the WEE virus was carried for 
17 successive insect-to-insect transfers without loss of titer (481). 
This virus can be cultivated in cultures of tissues from mosquito 
(650). The EEE virus also multiplies in the mosquito vectors (482). 

Although the virus multiplies in their tissues, the infected mosqui- 
toes do not show any recognizable symptoms of disease. This situa- 
tion is different from that of typhus rickettsiae in the body louse, 
which dies because of massive destruction of its intestinal mucosa by 
the rickettsiae ( see chapter 19 ) . 

The reprodution of viruses in the insect appears to be the explana- 
tion of tbe long persistence of transmitting ability after a single infec- 
tive meal. For dengue, transmission by mosquitoes has been obtained 
as long as 174 days after feeding, (88) the longest period tested. 
Once infectious, an insect will probably remain so for life. Virus 
transmission from generation to generation through the egg is excep- 
tional; it is doubtful whether it occurs at all for yellow fever. St. 
Louis encephalitis virus can be transmitted for two generations through 
the eggs of the chicken mite, Dermanyssus gallinae (612). 

The fact that viruses that are generally considered to be parasites 
of vertebrates multiply in their arthropod vectors raises the problem 
of the significance of the host-insect cycle in virus reproduction. 
There is no evidence for a relation similar to that obtaining, for 
example, for malarial parasites or for rust fungi, in which alternating 
forms of reproduction take place in different hosts, so that the host 
cycle reflects the reproductive cycle of the parasite. Viruses that are 
transmitted by insects in nature can also be transmitted directly 
from a vertebrate host to another by injection or other means, and, 
as we have seen, from insect to insect. All available facts can be 
explained by postulating that the passage of a virus through a vector 
is not an essential part of the life cycle of the virus and that insect 
transmission is simply the expression of the ability of one particular 
virus host to inoculate the virus successfully into another host. 

Vectors and ecology of animal viruses. The vector-virus relation 
should be considered from the standpoint of virus survival and, there- 
fore, of the distribution and the epidemiology of virus diseases. Let 
us consider WEE (549). This disease is sporadic in man and spo- 
radically epidemic in horses. In horses, it appears in widely scattered 
herds in the spring and subsides before fall. Mosquitoes, particularly 
Culex tarsalis, can transmit the disease to horses, but infection of mos- 
quitoes by feeding on infected horses has not succeeded, probably be- 

CH. 16 Vectors and Ecology of Animal Viruses 313 

cause of the low virus titer in the horse blood. It is almost certain 
that the natural pathway of the disease is not "horse-mosquito-horse." 
Moreover, adult mosquitoes do not survive the winter and do not 
transmit the virus through the egg. Instead, several fowl, both do- 
mestic and wild, are found to be infected in a symptomless way and 
can infect mosquitoes that feed on them. The fowl are infested by 
parasites, particularly mites, which can carry the virus through their 
eggs and can transmit it from chicken to chicken. The most likely 
cycle for this virus is, therefore, "fowl-insect-fowl," with horses and 
man as accidental hosts. This idea is in agreement with the theo- 
retical expectation of a mild, symptomless infection in the most stable 
reservoir. Some vectors may carry the virus through the winter, but 
their role does not seem to be important. 

An intriguing situation arose in connection with yellow fever when 
sporadic human cases were first recognized in jungle clearings of the 
South American interior, away from the usual coastal abodes of Aedes 
aegypti and in the absence of human reservoirs, which by that time 
had practically been eliminated. This "jungle yellow fever" was 
traced to infection of monkeys, with transmission from monkey to 
monkey or to man by mosquitoes other than Aedes aegypti. If the 
jungle yellow fever virus is introduced into an urban community, it 
can then be spread by A. aegypti (618). 

A similar jungle yellow fever is suspected to exist in central Africa, 
the presumptive cradle of yellow fever. Yellow fever is rare and mild 
among African Negroes, even if they are not vaccinated. This fact 
suggests a certain degree of selection for yellow fever resistance 
among the African Negroes during the long centuries of exposure to 
the virus. 

The examples given above illustrate the general causes underlying 
the geographical distribution of animal virus diseases and clarify some 
of the epidemiological problems in virus control. It is clear that 
neither human vaccination nor Aedes eradication, however successful 
in eliminating coastal yellow fever, can completely eradicate the dis- 
ease as maintained in the jungle. Similarly, vaccination of horses 
against WEE can protect these animals but cannot eliminate the virus 
harbored in its bird reservoirs. Vector control can go one step fur- 
ther. There is a great difference between these diseases and those 
caused by ubiquitous, presumably single-host viruses, such as measles. 
Vaccination, when possible, is the key to the control of the single-host 
diseases. Serological variation of the virus is probably the main 
source of complications in such control measures. 

314 Vectors and Survival of Viruses 

Another interesting consequence of the mode of transmission of a 
virus is its seasonal incidence. For example, in two concurrent 
summer epidemics of WEE and poliomyelitis (203), the WEE rapidly 
disappeared with the coming of colder temperatures, which caused 
rapid death of mosquitoes, whereas the poliomyelitis incidence contin- 
ued high until late in the fall. This observation speaks against any 
important role of insects in the spread of poliomyelitis. The causes of 
the summer incidence of poliomyelitis in man are unknown. We might 
speculate on a role of the summer vacation habits that bring together 
children from different areas, whose previous immunizing contacts, if 
any, may have been with serologically different strains of the polio- 
myelitis virus (492). 

Some complex relations of viruses to multiple hosts can be clarified 
by brilliant and painstaking work; a classic example is that of swine 
influenza (601). The disease is caused by a double infection with 
swine influenza virus and with the bacterium Hemophilus influenzae 
suis. It is epizootic in swine herds, where it appears regularly in the 
fall, although the virus does not persist long in pigs. Interepidemi- 
cally, the virus can be preserved for years in the lungworm, a nema- 
tode that infests the lungs of the pigs and there becomes infected 
with the virus. The lungworm retains the virus throughout its com- 
plete life cycle, which involves several stages of parasitism in the pig 
and in the earthworm. The pig becomes parasitized by the lung- 
worms after eating earthworms infested with the lungworms. The 
adult lungworms develop in the pig and transfer the virus to it. The 
virus cannot even be detected serologically in the lungworms during 
large portions of their life cycle ("masked" virus). The only proof 
of the presence of the virus is the fact that pigs that have ingested 
lungworms and have been inoculated with H. influenzae suis can 
come down with influenza. The actual outbreak of the disease is gen- 
erally brought about by any one of a variety of exposures to unfavor- 
able conditions ( wet, cold weather; poor hog house conditions, etc. ) . 
Prophylaxis may be aimed at controlling one or more of the links in 
the complex chain of events. 

Interestingly enough, swine influenza appeared as a new disease in 
1918, coincident with a world-wide outburst of pandemic influenza. 
It is possible that in pandemic influenza some bacterium may have 
played a role similar to that of the Hemophilus in the swine. 

Plant viruses. Insect transmission is the rule in most plant virus 
diseases. The specificity relations are similar to those observed with 
animal viruses. Some vectors are biting insects, for example, the 

CH. 16 Persistent Transmission 315 

vector of turnip yellow mosaic. However, the majority of the insect 
vectors of plant virus diseases are sucking insects, mostly aphids and 
leafhoppers, but also white flies, mealy bugs, and tingids. The tiny 
green aphid Myzus persicae, a common plant louse, feeds on a great 
many plants and transmits at least 50 different viruses of potato, 
bean, cabbage, onion, tulip, sugar beet, and other plants. The leaf- 
hopper Circulifer tenellus, the vector of the virus of sugar beet curly 
top, probably transmits only this virus and is the only species known 
to transmit it. In one instance, vectors belonging to different related 
genera are needed to transmit two barely distinguishable virus strains 
(81). Because of such specificities, single virus strains can often be 
isolated from mixed-infected plants by means of proper insects. 

Nonpersistent transmission. Two types of virus transmission by in- 
sects are observed in plants. These are the persistent and the non- 
persistent types (667). In the nonpersistent type, the insect can 
transmit the virus immediately after feeding and for a relatively short 
time thereafter (24 hours or less). This is true, for example, for potato 
virus Y and Myzus persicae. It was long supposed that transmission 
in these cases was simply due to mechanical transfer of virus by the 
contaminated mouthparts of the insect. The relation is apparently 
more complex, however (see 43). Nonpersistent transmission shows 
certain specificities. Aphids may fail to transmit viruses such as 
tobacco mosaic virus that are readily transferred mechanically, yet may 
transmit other viruses from the same host plant. Successful transmis- 
sion depends on the physiological condition of the insects, such as the 
degree of previous starvation. Viruses transmitted in a nonpersistent 
way probably do not have to be deposited into the plant phloem, 
whereas some viruses transmitted by the persistent mechanism prob- 
ably do. 

Persistent transmission. In the persistent form of transmission, the 
insect is usually not immediately infectious after feeding. The incu- 
bation period, which lasts from a few hours to several weeks, includes 
the time for passage of virus through the wall of the gut, into the 
blood, and into the salivary glands. For some viruses at least, there 
is actual multiplication of the virus in the vector. The viruses reach 
the new host with material injected by the insect's stylets. The stylets 
find their way, by an intercellular or intracellular path, through the 
epidermis and often reach the phloem tissue of the vascular bundles 
(figure 90). 

The passage of viruses through the wall of the gut has been demon- 
strated by Storey's classic experiments (632) on Cicadulina mbila, the 


Vectors and Survival of Viruses 

leafhopper that transmits the corn streak disease. This leafhopper 
exists in two strains, one that can transmit the virus, the other that 
cannot. The transmitting strain carries the dominant allele of a gene 

pair, whereas the nontransmitting strain is homozygpus recessive. 

Figure 90. Section through a potato leaf and an aphid ( Myzus pseudosolani ) 
showing the path of the insect's stylet reaching into the phloem. From: Dykstra 
and Whitaker, J. Agr. Res. 57:319, 1938. 

Storey showed that the nontransmitting insects can transmit the dis- 
ease if, after they have fed on infected plants, their abdomens are 
punctured. The puncture permits the passage of the contents of the 
gut into the blood. Thus the effect of the genetic difference appears 
to be at least in part a difference in gut permeability to the virus. 

The presence of virus in insect's saliva has been demonstrated by 
allowing leafhoppers infected with sugar beet curly top virus to feed 
on droplets of a sugar solution. The sugar solution receives virus 

CH. 16 Multiplication of Plant Viruses in Vectors 317 

and can give it to a new batch of uninfected insects that are later fed 
on the same droplets (124a; 609). The salivary glands of insects, 
however, often contain less virus than the blood (63). 

The deposition of viruses into the vascular tissue by their vectors 
cannot be the whole explanation of the success of insect transmission 
of plant viruses. Many viruses cannot be transmitted by needle in- 
oculation into the phloem, for example, aster yellows virus. It is pos- 
sible that the insect's stylets are less damaging than needles to plant 
cells, or that sometimes the virus may be combined with some in- 
hibitor present in the plant sap but not in the saliva. In some cases, 
a virus that affects the plant xylem is inoculated by its vector into 
this tissue (350b). 

Multiplication of plant viruses in their insect vectors. Several plant 
viruses transmitted by leafhoppers actually multiply in the body of 
their insect vectors, just as yellow fever virus multiplies in Aedes 
aegypti. The incubation period, especially when long, represents in 
part the time needed for virus reproduction. The evidence in favor 
of the multiplication of plant viruses in their insect vectors is now 
beyond question. 

Multiplication had long been postulated, for such virus as aster 
yellows, on the basis of the following observations (82; 389; 390): 
1. Insects fed on infected plants may remain viruliferous for a very 
long time, presumably for their whole life. 2. After feeding, the 
amount of virus recoverable from an insect decreases at first, then 
increases again. 3. This process can be exaggerated by heating the 
insects, a treatment that can render them nonviruliferous. If the heat 
treatment is properly administered, the insects may later become viru- 
liferous again, supposedly because of multiplication of residual virus, 

More direct and convincing evidence has come from experiments 
in which the viruses have been carried from insect to insect for such 
long series of passages that multiplication must have occurred for any 
virus to be present. At least two viruses can pass from mother tc 
progeny through the egg. Rice stunt virus was transmitted through 
the eggs of its vector for 6 generations (245), and clover club-leal 
virus for 21 generations (84). Had the clover club-leaf virus not mul- 
tiplied, it would have been diluted by a factor of at least 10 20 betweer 
the initial insect and the last batch of eggs tested. This result ij 
clearly incompatible with the hypothesis of dilution and can only be 
explained by multiplication of the virus in the insect. 

The discovery that some plant viruses can be transferred directl) 
from insect to insect by injection of extract of viruliferous insects intc 

318 Vectors and Survival of Viruses 

nonviruliferous ones (632) has made possible another proof of virus 
multiplication. Aster yellows virus has been transmitted for 10 pas- 
sages, corresponding to a dilution factor of 10 40 , in the leafhopper 
Macrosteles divisus (470), and the wound tumor virus has been 
transmitted for 7 transfers in the vector Agallia constricta (dilution 
factor 10 18 ; 85). In these experiments, reinfection through the 
plants on which the insects fed was excluded by using plants that do 
not support multiplication of the virus under investigation. Particles 
of the wound tumor virus isolated either from plant tissue or from 
insect tissue have the same morphology (lOOc). 

Although multiplication of plant viruses in their insect vectors has 
been proved for several viruses, it is not certain whether it occurs in 
many others. For example, curly top virus presumably does not mul- 
tiply in its vector, or t least does not multiply enough to maintain 
itself in insects that have acquired small amounts of virus. 

Whether many plant viruses multiply in the insects or not, there is 
no evidence for an obligatory alternation of plant and insect para- 
sitism. Both insect-to-insect transmission and plant-to-plant trans- 
mission appear to be possible in all instances that have been investi- 


Virus multiplication in the insect vectors of viruses parasitic on ver- 
tebrates and on plants raises a number of interesting problems. We 
must ask whether these viruses should be considered as plant viruses, 
insect viruses, or vertebrate viruses. It is clearly a matter of choice 
whether we call aster yellows virus a plant virus or an insect virus. 
The real problem concerns the origin of these viruses and their 
natural history. The primary host of a virus, that is, the host which 
is essential to its survival, is probably the one in which the virus 
produces the least damage and can best maintain itself alive from gen- 
eration to generation. Thus, insects in which a virus can be transmitted 
through the eggs, and which show no harmful effect of the virus, 
may be the most important reservoir. 

We know that virus reproduction requires a host possessing a spe- 
cific pattern of synthetic mechanisms, whose suitability for the virus 
is genetically determined. If a virus can grow in two hosts widely 
separated taxonomically ( such as a plant and an insect, or a mammal 
and a mite), it must find suitable conditions for growth in both hosts. 

CH. 16 Vectors, Host Range, and Evolution of Viruses 319 

In turn, this suggests that similar synthetic mechanisms have arisen, 
probably by different genetic evolution, in the two hosts. 

It is unlikely that an insect-transmitted virus, for example, aster yel- 
lows, could have remained parasitic on a long evolutionary series of 
plants and of insects feeding on them. It is reasonable to assume that 
the virus was originally restricted to one of the hosts only, the plant 
or the insect. The same may be said of insect-transmitted animal 
viruses. Suppose we started with a virus parasitic in an insect that 
feeds regularly on a plant, for example, the aster. Virus will regu- 
larly be inoculated into the plant, in which we assume it does not 
reproduce. If the virus happens to produce variants that can multiply 
in the plant tissue, whenever such a variant particle even by an 
extremely rare chance enters the plant, its chances of invading other 
insect individuals or even other insect species feeding on that plant 
are enormously increased. Such a variant will soon become the pre- 
dominant type. A mutant of a mammalian virus able to reproduce in 
(and reach the saliva of) one of the many types of mosquitoes that 
suck the blood of the mammalian host (particularly if the mosquito 
is not itself damaged by the virus) might spread so much better than 
the parent virus strain as to displace it or to occupy successfully many 
new ecological niches. 

Thus we may visualize the ability of a virus to multiply in two 
hosts, one feeding upon the other, as an adaptation, by natural selec- 
tion, to a mode of life more favorable for survival. The survival of 
an insect virus will be greatly favored by introduction into a longer- 
lived host, rendering it less dependent on egg transmission. In turn, 
the survival of a plant virus will be enormously favored by multipli- 
cation in an insect that feeds on plants. 

Although the paths of evolution of viruses are as yet difficult to 
trace, the above considerations suggest some of the possible features 
.of the process. Another source of information is provided by epide- 
miological data, as we have mentioned in connection with Western 
equine encephalomyelitis. Ecological considerations are also useful. 
They have been employed, for example, in attempts to trace the 
origin and evolution of tobacco mosaic virus and of the response of 
various species of Nicotiana to this virus (341a; 654c); in support of 
the origin of pox viruses of man and the domestic animals from an 
ancestor pathogenic for rodents; and of poliomyelitis from viruses of 
the mouse encephalitis group (114). Rodents, in view of their wide- 
spread distribution, their high fertility, and the continuous contacts 

320 Vectors and Survival of Viruses 

between wild rodents and their man-surrounding relatives, are prob- 
ably excellent reservoirs for many viruses of human diseases. 

In speculating along these lines we must remember that we are not 
only observing virus evolution but the evolution of virus susceptibility 
in the host populations as well. For example, the virus of measles 
causes a mild disease in human populations that have long been ex- 
posed to it, but it produced a highly fatal disease (with over 25% 
mortality) when first introduced among the inhabitants of South Sea 
islands. This suggests that the present virulence relation in the 
"civilized" world is due to the selection of resistant human hosts rather 
than to changes in virus properties. 

Let us state once again that the changes in host and virus popula- 
tions that are most favorable for virus survival are those that encour- 
age widespread and asymptomatic infection. Herpes virus infects 
about 90% of humans within their first 5 years of life and generally 
remains in the body in a symptomless relation. Other viruses related 
to herpes ( pseudorabies, virus B) also have a tendency to produce 
asymptomatic infections. The virus of silkworm jaundice, which can 
probably be transmitted through the insect egg, is carried in latent 
form by most silkworm colonies. A variety of stimuli can provoke 
the appearance of the disease ( 654a; 690a ) . 

A man-made experiment in virus ecology is the artificial introduction 
of myxoma virus in Australia as a means of controlling rabbits which 
have multiplied to enormous numbers in some parts of that continent. 
After several years of slow spread among the rodents, the myxoma 
virus, disseminated by contacts and with the help of some insects 
(215a) has suddenly spread widely in epidemic waves since 1951, 
destroying enormous numbers of rabbits (545a). 



Viruses and Tumors 

Tumors and cancers. The tendency of virus-infected cells to disin- 
tegrate, as for example in the lysis of phage-infected bacteria, is by 
no means general. Frequently, cells continue to divide after virus 
infection, as proved by the presence of mitotic figures in inclusion- 
bearing cells. The major manifestation of a virus infection may even 
be an abnormal proliferation of the host cells, which may lead to the 
formation of tumors. 

A variety of abnormal growths in multicellular organisms (animals 
or plants), in which a group of cells reproduces beyond its place in 
the organism, are called tumors or neoplasms. The cells that consti- 
tute a tumor are generally not different in any obvious gross way 
from those of normal tissues. Yet, the tumors are distinguished from 
pathological hyperplasia (an exaggerated growth of normal cells, as 
in the formation of excessive bone tissue in the callus around a bone 
fracture) by their lack of functional coordination with the rest of the 
body. This autonomy makes tumors truly parasitic on the rest of the 
organism. For example, a tumor of fat tissue or lipoma will not lose 
fat when the rest of the organism loses almost all its fat because of 
starvation or disease. 

The degree of parasitism upon the organism varies from tumor to 
tumor. Benign neoplasms are those that grow without actively de- 
stroying neighboring organs or disrupting the metabolism of the or- 
ganisms, and with no tendency to produce metastases, that is, sec- 
ondary tumors caused by transfer and implantation of tumoral cells 
into other organs. A typically benign tumor, such as a lipoma or a 
fibroma, is mainly a mechanical nuisance, as in growing it pushes 
aside other organs. The transition is almost continuous from benign 
to malignant tumors, the cancers of current speech. Cancers are neo- 


322 Viruses and Tumors 

plasms that grow very rapidly, whose cells exhibit frequent and often 
abnormal mitoses, infiltrate and destroy normal tissues and organs, and 
produce metastases. 

The changes that make cells neoplastic and, more especially, malig- 
nant are not yet understood. Usually tumor cells are related to nor- 
mal cells and may perform some of the specific activities, secretory 
or otherwise, of the tissues from which they derive, although without 
coordination with the normal body constituents. Thus^ve distinguish 
epithelial and connective tumors. Among the epithelial we find, for 
example, papillomas or epidermal tumors (warts), adenomas or 
glandular tumors; among the connective, we find chondromas or car- 
tilaginous tumors, osteomas or bony tumors, and many others. Even 
in the most atypical cancer, the tissue organization often bears some 
resemblance to that of a normal tissue. This is evident both in epi- 
thelial cancers or carcinomas ( in which the tumoral epithelial cells are 
accompanied by a nontumoral connective stroma), and in connective 
tissue cancers or sarcoma&f The tumoral quality of the cells is an in- 
trinsic property. In tissue cultures, a line of tumor cells maintains its 
tumoral properties and, if the cultivated cells are properly reimplanted 
by graft into a suitable host, they will "take" and produce a tumor. 

Transplantability by graft is one of the major techniques of experi- 
mental tumorgresearch. Tumor grafting is subject in animals such as 
mammals to the same type of limitations as the grafting of normal 
tissues. Genetic factors prevent the "taking" of heterologous grafts, 
partly because of immediate toxic reactions, partly because of the 
formation of antibodies against the foreign tissues. Heterologous 
grafts, even beyond the borders of the species and the genus, may 
succeed in places, such as the brain and the anterior chamber of the 
eye (275), or in chick embryos (497), where antibody production and 
penetration are restricted and other resistance mechanisms are prob- 
ably less effective. 

Graft experiments with tumors in laboratory animals generally re- 
quire highly inbred stocks obtained by repeated brother-to-sister 
matings (pure lines). An interesting manifestation of the relative 
autonomy of tumor tissues is that the genetic conditions for successful 
implantations are often somewhat less stringent than for most normal 
tissues; for example, some mouse tumors can be propagated by graft 
in many individuals of noninbred colonies (see 425). Moreover, extra- 
chromosomal factors appear to play a role in determining susceptibility 
to tumor grafting (141; 406). 

CH. 17 Etiology of Tumors 323 

In typical instances of tumor graft, it is the transplanted cells that 
give rise to the tumor in the new location, without neoplastic trans- 
formation of host cells. Following heterologous grafts, serological tests 
indicate that the transplanted tumor cells still possess the specificity 
of the donor. There is usually no reason to suspect a turnoral trans- 
formation of the cells of the recipient host by the material in the 
graft. The frequent appearance of sarcomas in the stroma of trans- 
planted carcinomas or even of benign adenomas is unexplained, how- 
ever; it may be due to a tumoral nature of some of the connective cells 
already included in the first transfer, although the spread of cancer- 
producing agents from one type of cell to another cannot be excluded 


Carcinogenic agents. What causes a cell or group of cells to assume 
the neoplastic habit of growth and to become a tumor? Generally, 
the cause is unknown; we speak of "spontaneous tumors." Within 
recent years, however, it has become clear that a number of agents can 
induce the formation of neoplasms. Notable among them is a group 
of cyclic hydrocarbons (for example, methylcholanthrene, 3,4-benzo- 
pyrene, 1,2,5,6-dibenzanthracene; see 276), whose application to nor- 
mal tissues provokes a high incidence of tumors. These chemically 
induced tumors include most varieties that occur spontaneously and 
also some that would not be observed otherwise, possibly because their 
spontaneous incidence is too low. Following inoculation into a rat, 
for example, of one of these carcinogenic substances, a series of tissue 
disorders occur, which at first are not very dissimilar from a chronic 
inflammation, but which lead more or less rapidly to the formation 
of a true neoplasm, often a cancer. In the neoplasm the carcinogen 
is no longer present, but the cells have become permanently modified, 

The carcinogenic hydrocarbons are the best example of a variety of 
agents, including radiations (615), parasites (see 161), and many 
chemicals, that act as provocative causes of cancer, by inducing or 
favoring a more-or-less irreversible neoplastic transformation of the 
cells, after which the agent appears to play no further role?) 

Virus tumors. There are some carcinogenic agents, however, that 
remain present in the tumors they induce and apparently continue to 
play a role in the development of the neoplastic cells, accompanying 
them, so to speak, in their growth career. These persistent and actuat- 
ing carcinogens (563) are viruses. 

324/ Viruses and Tumors 

In 1903 Borrel (95) put forward the suggestion of a possible virus 
origin of cancer. His evidence was mainly a comparison of the prop- 
erties of carcinomas with those of a series of virus diseases of epi- 
thelial cells (infectious epiihelioses, such as pox diseases and foot- 
and-mouth), in which cell proliferation was observed side by side 
with cell necrosis. Borrel pointed out that the causes of tumors were 
probably multiple. Valid proof of the viral etiology of a tumor would 
require the demonstration of the successful production of tumors by 
inoculation of cell-free extracts of other tumors into normal hosts (97). 
In 1911 Rous produced such proof. 1 

Fowl tumors. Rous succeeded in reproducing indefinitely by serial 
inoculations of cell-free, bacteria-free filtrates of tumor extracts a sar- 
coma that had appeared spontaneously in a Plymouth Rock hen and 
had been transplanted into several hens of the same race (562). The 
agent in the extracts gave rise to tumors similar to the one of origin; 
they appeared very fast, from 7 to 10 days after inoculation. This is 
a much shorter time than is required for the action of chemical car- 
cinogens. The agent of Rous sarcoma was also present in the blood 
of tumor-carrying chickens. It was shown later that the active agent 
is in the form of particles, about 70 m/* in diameter, with a ratio of 
particles to infectious units of about 2000 in the best preparations 
(139). Turnprs could be induced by injecting filtrates either into the 
connective tissue of chickens or into the blood. The agent was indeed 
a virus in the accepted sense of the word. 

Following Rous' discovery and an almost simultaneous similar one 
(244), (several viruses were isolated from a variety of tumors of chick- 
ens, ducks, and other birds and from cases of fowl leukemia (228). 
Although no viruses have as yet been isolated from fowl carcinomas, 
it seems possible (229) that for every fowl tumor there will ultimately 
be found an etiologically responsible virus. Often a virus can be ob- 
tained by extraction from the original tumor, without previous trans- 
plantation^ 285 ). 

Individual viruses of fowl tumors differ from each other; some of 
them are apparently as little related to one another as any two animal 
viruses taken at random. In general, each virus gives rise to tumors 
with the same histopathological features as the tumor from which it 
has been isolated, and only if introduced in the proper tissue. This 
discriminating, directional carcinogenesis clearly differentiates these 

1 Transmission of fowl leukemia by cell-free extracts and blood serum had been 
reported in 1908 ( 204a ) , but the neoplastic nature of this disease had not yet been 
recognized clearly. 

CH. 17 Etiology of Tumors 325 

viruses from the chemical carcinogens, which bring out the neoplastic 
potentialities of a variety of tissues in a variety of hosts. Clearly, the 
specificity relations observed with tumor viruses resemble the relations 
of specificity between viruses and host cells in general. 

/Most fowl tumors, in spite of their inducibility by viral agents, are 
not contagious under natural conditions. Animals kept in the same 
cage with cancer-bearing individuals do not show any higher inci- 
dence of tumors than control animals. No insect vectors have been 
found, and the problem of the normal transmission of these viruses 
remains unsolved. The agent responsible for fowl lymphomatosis is 
transmitted from hen to chick through the egg and possibly also 
through the sperm) (107). 

There was initially a great resistance among cancer workers against 
recognizing the virus tumors of fowl as true tumors, or their agents 
as true viruses. The suggestion was put forward that the virus pro- 
vided only the stimulus for the tumoral transformations of the cells. 
The virus, however, continues to reproduce and to play a role in the 
neoplastic development of the cell line derived from the initially in- 
fected cell. Many of the arguments whether the Rous agent and other 
tumor agents of fowl are true viruses or transmissible inducers of cell 
transformations (transmissible mutagens; 140) are in reality applicable 
to viruses in general. They do not set aside the fowl tumor agents, 
which at any rate qualify for inclusion among viruses according to the 
definition of virus adopted in this book. The arguments have rather 
to do with the mode of action of viruses in their host cells and with 
their relation to other cell components. 

J3ther virus tumors.) The fowl tumors did not Ions remain the only 
known examples of tumors produced by viruses, and me now recognize 
virus tumors in insects (78a), in amphibia, in mammals, and in plants^) 

\A frequent carcinoma of the kidney of leopard frogs is caused by a 
viruslike agent (429; 560)) Its cells contain peculiar intranuclear 
inclusions such as many viruses produce. 

^n mammals, viruses have been isolated from tumors of man, rabbit, 
and mouse. Human warts (skin and laryngeal papillomas) and the 
genital papillomas (condyloma acuminatum) are caused by a virus or 
a group of closely related viruses (216). The role of the virus in the 
occasional cancerization of genital warts is unknown) (see 563). 

\In rabbits, there occurs a group of virus diseases (nbroma-myxoma) 
intermediate between inflammations and tumors. The fibroma (597) 
is a mild proliferative disease of the subcutaneous tissue, endemic in 
wild rabbits, which regularly undergoes spontaneous regression. The 

326 Viruses and Tumors 

myxoma (576) is a highly fatal, more inflammatory process, which also 
involves some proliferation of the infected cells. A variant form of the 
fibroma virus produces a form of inflammation without extensive cell 

Another virus which causes oral papillomatosis in rabbits (513) is 
apparently present in most individuals of certain rabbit colonies, where 
it manifests itself only by causing the formation of specific antibodies. 

Figure 91. Papillomas produced by artificial inoculation of papilloma virus 
(Shope) in wild cottontail rabbit. Courtesy Dr. R. E. Shope, Rockefeller Institute, 
New York. 

Rabbit papilloma. A very interesting virus tumor of rabbits is a 
papilloma (598) frequently observed in the wild cottontail rabbits 
(figure 91). The virus extracted from the papillomas induces an epi- 
thelial growth intermediate between benign warts and malignant car- 
cinomas, with some tendency to full cancerous transformation (373). 
The tumor is easily reproduced by rubbing its cell-free extracts on 
scarified skin and is probably contagious among cottontails in nature, 
as suggested by its geographic distribution. 

A remarkable feature of the rabbit papilloma virus is that its inocu- 
lation into the domestic rabbit (an animal belonging in a different 
genus from the wild rabbit) results in the formation of papillomas that 
are similar to those of wild rabbits, but from which little or no infec- 
tious virus can be isolated. No tumor-producing agent is present in 
extracts of these tumors, nor are physical particles resembling those of 

CH. 17 Etiology of Tumors 327 

papilloma virus found in amounts similar to those in the papilloma 
of the wild rabbits (57). If the papillomas in domestic rabbits are 
protected from mechanical injury (which otherwise causes their hard, 
horny mass to fall off) they regularly give rise to very malignant 
cancers in which no infectious virus is observed (see 564). 

Is the virus absent from these tumors? Not quite. Serological data 
indicate that the virus persists in a masked, noninfectious form (602). 
The blood of domestic rabbits that carry a papilloma (or a carcinoma 
derived from it) reveals the presence of antibodies that neutralize and 
fix complement with the virus isolated from cottontail warts. (Extracts 
of the domestic rabbit papillomas injected into rabbits induce an im- 
munity against the virus.J This serologically specific virus "remnant" 
cannot be eliminated by grafting the cancers that arise from the papil- 
lomas through several animals hyperimmunized against the virus 
(371). Such a procedure would eliminate foreign viruses, which 
accidentally or experimentally may contaminate grafted tumors (557). 
It should be noted that antivirus antibodies in high titer do not pre- 
vent growth of grafted cancer cells, although they prevent not only 
infection by the virus but also the "taking" of grafted papilloma cells. 

These observations seem to indicate that the virus, although masked, 
is a persistent component of the tumoral cells of domestic rabbits, and 
probably one directly concerned in determining the tumoral properties 
of these cells. 

There is more evidence of the actual existence of virus, possibly in 
modified form, in domestic rabbit papillomas. It is occasionally pos- 
sible to demonstrate the virus in filtrates of extracts of these papillomas 
by injecting the filtrates into skin rendered hyperplastic by nonspecific 
irritants (241). Moreover, in two cases (590; 600) serial transmission 
in the domestic rabbit for 19 and for 14 passages has been successful. 
Return of the virus to the cottontail rabbit for 1 passage gave rise 
immediately to a strain of the "cottontail type," nontransferrable in 
the domestic rabbit (590). 

It is interesting to compare these findings on the papilloma virus 
and on the tumors it produces in its various hosts with the findings on 
bacteriophage in various host situations (see chapter 9). The masked 
papilloma virus, like the prophage in the lysogenic bacteria, is not 
present in infectious form. It can be detected serologically, whereas 
the prophage has not yet been detected in this way, possibly because 
of different antigenic organization in the two types of viruses. The 
masked papilloma virus presumably impresses certain specific prop- 
erties on the cells that carry it, like the lysogenic prophages. Oc- 

328 Viruses and Tumors 

casional production of active virus from the masked papilloma virus 
(the equivalent of prophage maturation) may be hard to detect in 
view of the rather insensitive tests available for detection of papilloma 
virus ( 1 infectious unit = several millions of virus particles ) . Irradia- 
tion of domestic rabbit papillomas with x-rays often causes an increase 
in the amount of infectious virus detectable in extracts made a few 
hours later (240), an observation that recalls the "induction" of lyso- 
genic bacteria by radiations (454). 

One of the cancers derived from domestic rabbit papilloma, the Vx2 
carcinoma, retained the antigenic remnant of the virus through the 
first 22 transplants but "lost" it sometimes before the 46th transplant 
(566). In this case, the virus might have completely disappeared; or, 
it might have become reduced to a true "provirus," without detectable 
serological cross-reaction with the mature virus. The question would 
then arise as to how such provirus could be detected, a question we 
shall discuss again in a later section. 

Another interesting comparison is one between the masked papil- 
loma virus and that of phages phenotypically modified by their hosts 
(see page 295). The cells of domestic rabbit might modify the virus in 
such a way that it was not infectious for either domestic or wild rabbit, 
while remaining potentially infectious for some other host. Such a 
possibility would be difficult to test in the specific case of papilloma. 
The domestic rabbit papilloma does not contain any important amounts 
of particles similar to the virus particles; this fact indicates that the 
virus, if present, must be in a thoroughly modified form. 

The milk factor. Mammary cancer in mice has long been a favorite 
material for grafting studies and for the analysis of genetic factors in 
cancer causation. In 1936, Bittner (79) discovered that the tendency 
to develop mammary carcinoma was transmitted from the mother to 
the young through the milk, and that a "low-incidence" strain of mice 
could by foster-nursing be made into a "high-incidence" strain, and 
vice versa. Genetic factors are involved in determining the proportion 
of realized cancers and the persistence of the influence or milk factor 
in the milk of successive generations. 

This milk factor exhibits all the properties of a virus (23), including 
stimulation of specific antibodies. It is present not only in the milk 
but also in many tissues of the animal. In the cancers it is found in 
very large amounts, and can be seen in electron micrographs of cancer 
cells cultivated in vitro (figure 92; 527). The virus is not transmitted 
through the placenta, since Caesarean delivery from high-incidence 
mothers yields animals with low cancer incidence. To establish itself, 

CH. 17 

Etiology of Tumors 


the virus must enter very young mice, apparently before the wall of 
the intestine becomes impermeable to it. This is not the only reason, 
however; direct injection of milk factor in adult mammary tissue gen- 
erally fails to transmit the cancer. The cancer only develops in adult 

Figure 92. Electron micrograph of a portion of a cell from a tissue culture of 
mouse breast carcinoma. Note the abnormal, presumably viral, particles. The 
large, dark masses are mitochondria. From: Porter and Thompson (527). Cour- 
tesy Dr. K. R. Porter, Rockefeller Institute, New York. 

females that carry the milk factor and that have given birth to one or 
several litters, unless artificial hormonal stimuli similar to those of 
pregnancy are provided (393). 

Thus, the milk factor enters in early life, localizes itself in various 
tissues of the host, including the mammary glands, and remains there 
latent for most of the host's life. It produces tumors (benign at first, 
but becoming malignant ) only in mammary gland cells that have been 

330 Viruses and Tumors 

conditioned by hormonal influences. It is difficult, here, to decide 
which is the provocative and which the actuating cause of the cancer; 
certainly, the role of the milk factor in breast cancer is more deter- 
mining than the role of foster nursing in the transplantability of cer- 
tain tumors (141). We cannot exclude the possibility that different 
forms of the milk-factor agent may exist, some determining high inci- 
dence, some low incidence of cancers, and that one form may mutate 
to the other. Such a possibility could explain sudden changes to 
high incidence in low-incidence lines of mice that had been obtained 
by foster nursing (80). 

A claim (644) that, from mammary cancer of mice cultivated in 
the yolk sac of the fertile egg, a virus directly carcinogenic to mice 
by injection could be recovered has not been confirmed (652). In 
mouse leukemia, an agent analogous to the milk factor may be in- 
volved. Cell-free extracts from leukemic . cells, used to inoculate 
1-day-old mice, may cause leukemia (279). 

Virus tumors of plants. In plants, where most tumors (galls) are 
of bacterial or fungal origin, a wound tumor virus has been described 
(83). It causes a systemic disease, part of which consists of the for- 
mation of multiple tumors on roots and stems. The tumors develop 
especially at the site of wounds artificially produced or at the places 
where the tissues are cracked by the emergence of lateral roots or 
shoots. The virus is insect transmitted and has a wide host range. 


In all the tumors discussed above there is excellent reason for 
implicating viruses as causative agents. We may then ask: What is 
the role of the virus in transforming normal cells into tumoral cells? 
Does it act in the initial stage only, or throughout the development and 
growth of the neoplasms? 

The tumor viruses reproduce with and inside the tumor cells, and 
the cells cannot be freed of the active virus, except for special cases 
of masking. The problem, thus, is one of the respective roles of the 
viral and of the cellular components of an integrated system, the 
virus-infected cell, in determining the neoplastic properties. As such, 
this problem and the more general problems of the roles of various 
cell components in cellular differentiation are still unsolved. 

Stimulation of cells to increased reproduction as a result of virus 
infection is not infrequent. We observe it in many "destructive" virus 
infections, and a whole series of viruses can be arranged in order of 

CH. 17 The Role of Viruses in Tumor Growth 331 

increasing stimulatory action and decreasing necrotizing action (table 
31). Tumor viruses themselves may cause necrotic reactions; for 
example, the Rous sarcoma virus causes an acute hemorrhagic disease 
in very young animals (197). 

Table 31. From necrosis to neoplasia 

From Duran-Reynals and Shrigley (108) 

Epithelial Series Connective Tissue Series 

Foot-and-mouth disease Canary pox 

Vaccinia Vaccinia 

Fowl pox and sheep pox Infectious myxoma 

Filtrable warts (human) Infectious fibroma 

Infectious papilloma (wild rabbits) Myxosarcomas of Rous type 

Nonfiltrable papilloma (domestic rabbits) Fibrosarcoma of chickens 

Basal cell carcinoma * Nonfiltrable fowl tumors * 

Malignant epithelioma * Mammalian sarcoma * 

* Tumors without known relation to viruses. 

It was pointed out by Rous (564) that tumoral manifestations follow 
intracutaneous injection of fat-soluble dyes such as Sudan HI; but 
these manifestations regress completely with resorption of the dye, as 
though the stimulus had to be present all the time for continued tumoral 
proliferation ( 222 ) . In virus tumors, the situation might be analogous, 
virus reproduction providing for the persistence of the stimulus. 

There is no evidence that the neoplastic character of the cells of 
tumors formed in response to tumor viruses can be maintained in the 
absence of virus. Cells from virus tumors of fowl or rabbit are as 
transplantable as those of other tumors, but the virus remains harbored 
in them. The transplanted virus tumors of fowl grow mainly by 
reproduction of the implanted virus-carrying cells, rather than by 
neoplastic transformation of host cells, and the same is true of the 
metastases of the virus tumors. This, however, seems a reasonable 
course of events, since the virus is not necessarily liberated by the 
transplanted cells, which actually shelter it from host antibodies. 
Even when virus is liberated, the probability that it will cancerize 
other host cells may be as low as in the admittedly difficult induction 
of these tumors by injection of cell-free extracts. It is possible that 
the virus itself, and not only tumor cells, is sometimes responsible for 
the formation of metastases of fowl sarcomas, since the virus does 

332 Viruses and Tumors 

circulate in the blood. Metastases, probably due to virus, have been 
obtained after fowl sarcomas had regressed as a result of x-ray treat- 
ment (515). 

As for the role of the virus in determining the characteristics of a 
tumor, we have already mentioned its evident importance; for example, 
a virus from an osteochondroma of chicken causes formation of bone 
and cartilage from connective tissue cells if injected in other chickens 
(564). The host cells must be of the proper type and age. The cells 
that under virus action give rise to Rous sarcoma, for example, are 
thought to be mainly macrophages (123) or fibroblasts (430). Only 
rarely can a virus cause formation of a new type of tumor by infecting 
a different type of cell, and this may involve virus mutation and selec- 
tion, as in a sarcoma produced by fowl leukomatosis virus that had 
been stored in glycerin ( see 501 ) . 

Rous sarcoma and other fowl tumors will "take" more easily in 
young chickens than in older ones. Inoculation into other species, such 
as duck, pheasant, and guinea fowl, succeeds better in very young birds 
than in adults, with some differences from virus to virus. We have 
here the exact parallel of situations known to obtain for most viruses, 
which require special conditions for transmission to hosts of other 

The neoplastic characteristics of the tumors formed in a heterologous 
species reflect the character of the virus, as shown by a number of 
observations. Injection into ducklings of filtrates of Rous sarcomas 
from chickens, when successful, causes relatively early formation of 
tumors that often regress, are poorly transplantable, and are still 
"chicken tumors," since extracts from them, if reintroduced in chickens, 
give rise to typical sarcomas with normal frequency. Transplantation 
of the "heterologous" tumors can be prevented by the use of antiserum 
against chicken tissue, an indication that the virus has actually con- 
ferred to the tumor developing in the duck some species specificity of 
the chicken (197 a). Similarly, Fujinami sarcoma of ducks shows 
"duck tumor" characters (537, 538). Sometimes, the Rous sarcoma 
virus from chicken tumors, upon inoculation in ducks, gives rise to a 
different type of tumor, which develops late and is more easily trans- 
missible from duck to duck. The virus from these late tumors has 
lost its typical affinity for chicken tissues, a fact that suggests a virus 
mutation (or host-controlled transformation) from "chicken type" to 
"duck type" (197a). 

The situation with amphibian tumors (560) closely parallels that 
with virus tumors of fowl, and provides an even better illustration of 
the interplay of virus and host properties in determining not only the 

CH. 17 Viruses and Spontaneous Tumors 333 

characteristics of neoplastic growths but also the characteristics of the 
viral agents obtained from these growths. Graft of the virus carcinoma 
of the kidney of Vermont frogs (429) into limbs of newts led to the 
formation of cartilaginous tumors in other limbs of the newts. From 
these cartilage tumors an agent could be obtained which, reintroduced 
into the Vermont frog, produced, besides the renal tumor, also bone- 
cartilage tumors of a type never otherwise observed in these frogs. 
Material from these tumors continued to produce both bone tumors 
and renal tumors by proper inoculations; it could also produce de- 
generative, histolytic changes, especially in muscles. Moreover, im- 
plantation of the bone tumors of newts into Wisconsin frogs induced 
renal cancers in the latter, whereas direct implantation of renal tumor 
from Vermont frogs to Wisconsin frogs failed to do so. 

A remarkable feature of the bone tumors induced by derivatives of 
the frog virus is the high frequency of the formation of two tumors in 
exactly symmetrical bilateral locations, for example, "at the joint be- 
tween the third and fourth phalanges of the fourth toe of both feet" 
(560). We gather the impression that the carcinogenic potentialities 
and the host-cell affinities of the virus are directly modified and con- 
trolled by subtle developmental properties of the host cells in which it 
has been produced. 

The suggestion was made (56'0) that these transformations of the 
viral agent may be induced by some form of genetic recombination 
between the agent and some tissue-specific, cytoplasmic particles sup- 
posedly involved in the determination of the specific properties of 
various tissues of the host. More generally, and without introducing 
hypothetical entities into the picture, we may visualize the possibility 
of host-controlled alterations of the host-range and morphogenetic 
potentialities of a virus, along the same lines as the host-controlled 
phenotypic variations of bacteriophages (76; 446; see page 295). 
These alterations might well reflect subtle developmental differences* 
of the host cells. 

Be this as it may, we conclude that there are excellent grounds for 
attributing an essential role to the tumor viruses in the determination 
of the reproductive and neoplastic activities of their host cells. 


The evidence given in the preceding section emphasizes the role of 
the virus in determining the properties of a virus-induced tumor. At 
the same time, it indicates a very intimate fusion of cellular and viral 

334 Viruses and Tumors 

properties. This raises the question: How far can this fusion go? May 
there not exist, in other tumors, viruses or viruslike agents, whose 
presence is responsible for the tumoral reproduction but which cannot 
be isolated in infectious form because their relation with their host 
cells is too intimate? 

The large majority of neoplasms is not transmissible by cell-free 
extracts. Since the number of individual tumors from which viruses 
have been isolated is only a small fraction of those tested, and since in 
all probability the great majority of mammalian tumors, if tested, 
would yield no virus, one may be led to thinking that the virus tumors 
are the exception, and that viruses have nothing to do with the causa- 
tion of most neoplasms. 

Most pathologists incline to the opinion that the causes of tumors 
are multiple, and that the neoplastic transformation represents a re- 
sponse to any one of a number of cell disturbances. Thus, chronic 
irritations would favor cancerization by providing a stimulus to cell 
reproduction that may go beyond the needs for tissue repair. Another 
widely entertained hypothesis is that most spontaneous tumors arise 
by somatic mutations, that is, by genetic changes in some somatic cells, 
which transform them into neoplastic cells, just as a genetic change in 
a cell of the reproductive line may cause the appearance of a new 
hereditary teait. Another explanation of the origin of some spon- 
taneous tumors considers their cells of origin as embryonal remnants, 
which late in life manifest their undiminished growth potentialities 
by giving rise to tumors. Tumors (teratomas) stemming from islets 
of embryonal tissues are, indeed, well known. 

The question of a possible virus etiology of apparently spontaneous 
tumors cannot be summarily dismissed. Since the days of Borrel, the 
virus theory of cancer has had a growing body of proponents. It is 
undeniable that the virus tumors are the only neoplasms whose direct, 
actuating cause is understood. It is only reasonable to follow a lead 
that may direct us toward a unified theory of tumor etiology. More- 
over, the impressive histological and biological resemblance between 
virus tumors and "nonvirus tumors" suggest a basic similarity of their 
etiological mechanisms as well. Intranuclear inclusions similar to those 
that characterize some virus-infected cells have been observed in 
tumoral cells, for example, in human gliomas (569), although they 
may have been due to viruses accidentally infecting the tumors (25). 

Faced by the lack of transmissibility of most tumors by cell-free 
extracts, the proponents of the virus theory have turned to the hy- 
pothesis of latent and masked viruses as the causes of most cancers. 

CH. 17 Viruses and Spontaneous Tumors 335 

Whatever the ultimate outcome of this hypothesis (the case for which 
has ably been presented by Andrewes (25) and by Oberling (501), 
among others), it must be said that it has proved a fruitful working 
hypothesis. It has stimulated an impressive body of experimental work 
and has directed the thinking of pathologists toward fundamental 
biological problems. 

The virus theory of neoplasms. As a starting point, let us consider 
again the rabbit papilloma of Shope (figure 93). The virus extracted 

papilloma in cottontail hare 


26 42 

papdloma papdloma aging . carcinoma transplants . carcinoma tfans P |ants 1 carcinoma 

ex,ra, jextrac, 

virus papilloma papilloma papilloma no antigen 

antigen antigen antigen (no virus? 

(masked provirus?) 

Figure 93. Diagram of the transformations of rabbit papilloma virus in various 

from cottontail papillomas, if injected into domestic rabbits, causes 
tumors, which are papillomas at first and later become cancerized. 
From these tumors little or no infectious virus can be isolated. The 
domestic rabbit tumors contain an antigen related to the virus by its 
ability to stimulate production of antivirus antibody. The virus, ac- 
cording to Shope's terminology, is masked (602) in the domestic rabbit 

Could the masking be due to a combination of virus with inhibitors 
(possibly antibodies)? There are cases, for example, of fowl tumors, 
from which active virus can only be obtained after removal of virus 
inhibitor by differential centrifugation (136). An inhibitor for the 
papilloma virus is actually formed (598) in domestic rabbit papil- 
lomas; yet, inhibitors, and especially antibodies, are certainly not the 
main cause of masking. Extraction methods that reveal large amounts 
of specific virus material in cottontail papillomas fail to extract any 

336 Viruses and Tumors 

similar material from the domestic rabbit papillomas (57). Admit- 
tedly, the difference between domestic and cottontail rabbits might 
only be in the amount of. virus produced. The titration methods for 
papilloma virus are not very sensitive ( several million particles per in- 
fectious unit), and the domestic rabbit cells may not produce enough 
virus (241). Yet, it has been suggested that the viral material in the 
domestic rabbit tumors is in a different, noninfectious form, and we 
are reminded of the nonrecoverability of other viruses in the early 
stages of infection with most viruses ( see chapter 13 ) and in cases of 
persistent latency, as for prophage in lysogenic bacteria (chapter 9). 
We may suppose that the virus, although present in the domestic 
rabbit tumors, never "matures" into infectious particles, just as in 
lysogenic bacteria a phage very seldom reaches maturation. 

The incompleteness of the masked virus might be due to the fact 
that production of fully infectious virus particles requires some specific 
component available in the cottontail, but not in the domestic rabbit. 
There is some ground for such a "two-factor" hypothesis (284), at 
least by analogy. For example, Fujinami sarcoma virus grown in 
ducks is neutralized by serum against normal duck tissue, whereas the 
same virus grown in chickens is neutralized by serum against normal 
chicken tissues. Virus grown in either host is also neutralized by 
serum against virus-containing tissues of either host. The two anti- 
bodies are quite different: neutralization by the host-specific one re- 
quires complement; neutralization by the virus-specific one does not. 2 

The important point is that the induced papillomas of domestic 
rabbit and the cancers derived from them, to an observer unaware 
of their causation by papilloma virus, would appear indistinguishable 
from any spontaneous tumor. They are transplantable, nonfiltrable, 
and have a tendency to cancerization, just like other skin papillomas, 
either spontaneous or caused by chemical carcinogens. The presence 
of the masked virus in the tumors is detected only by testing the 
serum of the tumor-bearing rabbit for antibody against the virus 
from cottontail, but an unwarned observer would not even suspect 
the existence of such a virus. 

2 These observations (which should be confirmed by more extensive studies) can 
be explained independently of the two-factor hypothesis; for example, the antibody 
against normal tissue might remove from the virus a protective protein of cellular 
origin, in the absence of which the virus may become rapidly inactivated. 

CH. 17 Search for Masked Viruses 337 

The suggestion that viruslike agents are present and etiologically 
responsible for all neoplasms requires an answer to several questions: 
1. How can these agents be detected, so that the hypothesis can be 
tested? 2. Can the suggestion be reconciled with the known facts of 
carcinogenesis? 3. Where do the viruslike agents come from? 4. When 
and how do they enter the cells which they will make neoplastic? An 
attempt to answer these questions will be a useful introduction to a 
discussion of the origin and natural relationship of viruses. 

Serological search for masked viruses in tumors. Could we test 
for a "viruslike something" in spontaneous tumors by serological means? 
There have been many failures in such attempts, but some positive 
results have been reported. The Brown-Pearce carcinoma of rabbits 
yields a nucleoprotein antigen, whose specific antibody prevents the 
graft of the tumor cells and their growth in vitro (372). Since, how- 
ever, the Brown-Pearce tumor had stemmed many years before from 
an animal of unknown genetic constitution, and was transmitted by 
graft in rabbits of various genetic lines ever since, the immunity could 
be of the genetic-incompatibility type rather than a tumor-specific one. 

Suppression of tumor growth by antibodies against the tumor cells 
or against their extracts is by no means infrequent.. The cancer lit- 
erature contains a number of examples of tumors that after trans- 
plantation regress and leave the animal solidly immune to new grafts 
of the same tumor (see 129). Most such cases, however, can simply 
be interpreted as due to the heterologous character of the implanted 
cells, not to specific antibodies against the tumor cells as such. The 
serological evidence in favor of the presence of viruses in spontaneous 
tumors must be considered as quite inadequate. 

Yet, even if the serological search for viruses in spontaneous tumors 
should prove fruitless, the latent-virus theory would hardly be re- 
futed. Indeed, in the best-studied cases of virus latency, those of the 
prophages in lysogenic bacteria, the provirus has not been detected 
serologically. All available evidence suggests that most of the anti- 
gens of the mature phage are acquired only during maturation and are 
not present in the prophage (313). 

How can we exclude the possibility that tumors are caused by pro- 
viruses, with no easily demonstrable antigens, whose maturation may 
either fail completely or yield "virus" with little or no infectious power? 
Actually, the Vx2 carcinoma, which had originated from rabbit papil- 
loma, but does not now contain the virus-related antigen (566), may 
be an example of the complete evolution from a full-fledged virus 
(papilloma virus) to a masked virus (the antigen in domestic papil- 

338 Viruses and Tumors 

lomas) to the "pure" provirus. The question is: How could such a 
pure, nonmaturing provirus be meaningfully distinguished from any 
"normal" protoplasmic constituents? We shall discuss this question in 
chapter 18. 

The virus theory and artificial carcinogens. A reconciliation of 
the virus hypothesis with chemical carcinogenesis has been sought in 
a number of ways. Papilloma virus, injected intravenously into rabbits 
whose skin had previously been rubbed with tar or other carcinogens, 
localizes itself in the treated areas and causes a rapid formation of fast- 
growing cancers, definitely more malignant than the growths produced 
by carcinogen alone (370; 565). Tarring of the skin causes even the 
almost innocuous fibroma virus to give rise to less regularly regressing 
growths, and occasionally to generalized fibromatosis or even to true 
sarcomas (30). These neoplasms are transplantable and do not yield 
any virus. 

Thus, carcinogens can provide conditions that localize and enhance 
virus action. Chemical carcinogens could act through viruses, if 
viruses were present; what has to be explained is the supposed pres- 
ence of the viruses. Where could they come from, when not intro- 
duced deliberately in our experiments? 

Several puzzling lines of evidence are available. Mclntosh (456) 
produced sarcomas in chickens by injection of tar and found that 
some of these tumors could then be propagated by cell-free filtrates; 
other workers, however, could not duplicate his results (515). In 
more convincing experiments, it was shown that the blood of birds 
that carry tar-induced, nonfiltrable tumors often contains antibodies 
against various viruses, including Rous sarcoma virus (26). It should 
be pointed out that many of the fowl tumor viruses give some sero- 
logical cross-reactions among themselves. Thus the carcinogens may 
act by providing the conditions for the neoplastic action of latent 

The carcinogen might either produce a nonspecific tissue reaction, 
which could make the cells susceptible to the neoplastic influence of 
a virus (or provirus) already present within them, or it might act on 
the virus itself. The recently discovered ability of certain carcinogens 
to induce hereditary mutations in various organisms (176) has been 
considered as supporting the idea that tumors arise by somatic muta- 
tions, but it could equally well be used in support of the virus theory, 
by assuming that the carcinogens cause mutations in latent viruses. 
The somatic-mutation theory of cancer origin, besides lacking direct 
positive evidence to support it, must overcome the test of explaining 

CH. 17 Viruses and Artificial Carcinogens 339 

the well-known precancerous reactions, in which the cells, although 
modified, are not yet neoplastic. When does the mutation take place? 
Moreover, how can the somatic-mutation theory account for virus 
tumors? Some ingenious suggestions, such as the requirement for 
many independently arisen mutant cells to form a cancer focus (223), 
do not overcome the above objections. The virus theory of neoplasms 
might account for hereditary influences on the incidence of cancer 
(see 425), which may actually represent quantitative genetic differ- 
ences in virus susceptibility. A case could even be made for a role 
of somatic mutations in making a cell susceptible to carcinogenesis by 
a latent virus. 

As far as carcinogens such as parasites are concerned, their role 
could even be interpreted as that of carriers of tumor viruses in 
masked form. A rather good model is provided by the role of the 
lungworm in the transmission of swine influenza virus to the hog. 

A remarkable case is that of the crown gall tumor of plants (178; 
607). This tumor is produced in response to infection with the bac- 
terium Phytomonas tumefaciens. The tumor cells continue their neo- 
plastic growth, both in plants and in tissue cultures, after bacteria 
have been eliminated in any one of a variety of ways. The bacterium 
apparently unleashes the neoplastic potentialities of the host cells but 
is not needed for their continued expression. Still, a specific agent 
responsible for tumoral growth seems to be present in the cells of the 
crown gall tumor (101). When a fragment of tumor is forced to 
undergo exceptionally rapid growth, it ultimately gives rise to normal 
plant tissue, which can reconstitute normal plant organs and. become 
indistinguishable from original uninfected tissues. Forced rapid vege- 
tative growth is known to remove from cells of various organisms cer- 
tain cytoplasmic elements responsible for some of the properties of the 
organisms, such as the "killer" factor of Paramecium aurelia (530; see 
chapter 18). The abnormally rapid growth of the cells leads to dilu- 
tion and removal of self-reproducing elements, whose reproduction is 
apparently not synchronized with cell reproduction and cannot keep 
up with an increased pace of the latter. Loss of latent viruses by 
analogous processes, for example, the loss of prophage from lysogenic 
bacteria repeatedly transferred in media unfavorable to phage repro- 
duction, has been reported (135). 

It seems plausible that the neoplastic character of the crown gall 
cells is determined by a viruslike agent of some sort, which can be 
eliminated from rapidly growing cells. The mechanism of the initiat- 
ing role of the bacterium Phytomonas tumefaciens in crown gall tu- 

340 Viruses and Tumors 

mors remains obscure. It might be that the bacterium brings into the 
plant a self -reproducing viruslike agent ( possibly a bacteriophage ) ; or 
it might stimulate the activity of a latent virus; or it might transform 
some normal cell constituent into a tumor-producing agent. More in- 
formation will be needed before an answer can be given, and the 
ultimate answer may be quite different from any one now predictable. 

Latent viruses and tumors. Chemical, genetic, and parasitic car- 
cinogenesis, then, does not present unsurmountable objections to the 
theory that tumors are caused by latent viruses or viruslike agents. 
We must see if we can account for the postulated latent viruses them- 

There is a whole series of transition cases between latent viruses of 
unknown derivation, such as pneumonia virus of mice (486) or virus 
HI (558), and viruses that can be observed to become latent under 
controlled experimental conditions. Thus, the virus of lymphocytic 
choriomeningitis has a reservoir in the wild grey mouse ( 31 ) and can 
be transmitted to man and monkey by ticks; if introduced accidentally 
or purposefully in sC colony of white mice, the virus infects in a symp- 
tomless way practically all exposed animals, being transmitted to the 
young, at first by infection in early age, later even in the womb (651). 
Active virus can be recovered from many animals of the colony, and 
the presence of antibody reveals a past or present infection in the 
others. Injection of irritating substances, sufch as sterile broth, into 
the brain of virus-carrying mice can revive the pathogenicity of the 

The milk factor of mammary tumor in mice provides an example of 
a latency lasting from the time of infection (only possible in early 
age ) to the time of the appearance of properly susceptible cells in the 
breast of the multiparous females. The virus is not fully masked, since 
it can be demonstrated in many parts of the organism by feeding new- 
born mice with organ extracts; the mice later prove to be carriers of 
the factor (23). The latency seems to reflect only the lack of cells 
that can react neoplastically to the virus. A number of stimuli, such 
as hormones and carcinogens, can modify the breast cells in such a 
way that the tumors develop earlier. The virus, however, multiplies 
in a variety of tissues that do not show any abnormal response to its 

Finally, the facts of lysogenicity in bacteria, and especially of in- 
duced lysogenicity by infection of bacteria with temperate phages, 
provide us with a variety of conditions in which viruses become latent 
( = asymptomatic ) and masked ( = noninf ectious ) . 

CH. 17 Latent Viruses and Tumors 341 

Thus, virus latency and masking phenomena could provide condi- 
tions simulating the apparent spontaneous origin of tumors from which 
ho transmissible virus can be obtained. Should we then suspect the 
existence of an enormous variety of latent viruses, each responsible for 
one specific type of tumor? This may not be necessary. The neoplastic 
potentialities of each tumor virus are not very restricted; the same virus 
strain can give rise to different manifestations in different hosts and 
tissues. For example, sarcomatous growth has been produced by fowl 
leukomatosis virus; a variety of different sarcomas are produced by 
Rous sarcoma virus in ducks, pheasants, and other birds; several bio- 
logically different carcinomas have arisen from rabbit papillomas. We 
have already described (page 332) the remarkable situation observed 
with the amphibian tumors provoked by the agent of the renal car- 
cinoma of frogs ( 560 ) . It is clear that virus variability, either of the 
genetic type or of the host-controlled type observed with bacterio- 
phage, can well provide the necessary variety of tumor types from a 
relatively small number of latent viruses. 

Conversely, the potential variety of latent viruses might explain why 
a given carcinogen produces different types of tumors in animals of 
different strains and races; the neoplastic potentialities of the cells 
would reflect, at least in part, the nature of the viruses they carry. 

The real difficulty for the virus theory of tumors is the requirement 
for an almost ubiquitous distribution of latent tumor viruses. Where 
can these viruses come from? Simple transmission from tumor- 
bearing animals to healthy animals is out of the question, since even 
the experimental transmission of the known tumor viruses is difficult; 
the tumor-bearing animals would not represent an adequate source of 
viruses. Entry in early age (possibly in nursing or by vector trans- 
mission, or even in the intrauterine life through the egg) is apparently 
not frequent for viruses in general. We have already mentioned the 
failures to demonstrate milk-transmissible agents in many tumors, in- 
cluding some breast cancers of mice. The milk-factor tumor is appar- 
ently an exception rather than a prototype. 

Some lines of research provide suggestive information. Thus, many 
apparently normal chickens possess antibodies against Rous sarcoma 
virus and other fowl tumor viruses. These antibodies develop in 
chicks only after hatching. Their appearance can be prevented by 
isolating the newly hatched chicks from adult chickens; conversely, 
their frequency is increased by contacts with tumor-bearing chickens. 
The suggestion is that most infections with tumor viruses, just as those 

342 Viruses and Tumors 

with certain other viruses, are subclinical and nontumoral, and lead to 
virus latency; the tumoral transformation would be the exception. 

Even if a case could be made for an almost ubiquitous distribution 
of tumor viruses, several problems would still remain unanswered, for 
example, in connection with the transplantability of tumors. If latent 
viruses were ubiquitous, what role would they play in the serological 
conditioning of graft "taking"? Why would the animals of a highly 
inbred line not carry the same latent viruses, and therefore be immune 
against tumor transplantation? Why should a spontaneous or chemi- 
cally induced tumor of one of these animals ( or a virus derived from 
it, as in fowl sarcomas) find it easier rather than harder to "take" in 
animals of the same breed? Moreover, agents such as Rous sarcoma 
virus can cause the neoplastic transformation of normal cells infected 
from outside; if they were ubiquitous, why would they not cause it 
regularly from inside? Could the true tumor viruses be mutants of 
ubiquitous, less active agents? 

To many of the adherents of the virus theory of tumor etiology the 
idea of tumor viruses as latent exogenous parasites appears unsatisfac- 
tory. A role of exogenous viruses can hardly explain the specific neo- 
plastic potentialities of different embryonal tissues transplanted into 
adult mice of the same breed, a situation that would require early 
intrauterine infection with a variety of different viruses ( 564 ) . Some 
of the workers in the cancer field have gone a step further, and have 
suggested that tumors are indeed caused by viruslike agents, but that 
these are new viruses, that is, normal components of cells, which under 
the influence of any one of a variety of stimuli, or even by spontaneous 
mutation, have embarked on an abnormal career, conferring a neo- 
plastic habit to the cells that harbor them. In extreme cases they 
could become transmissible from cell to cell and therefore become 
true viruses. Support for this viewpoint has been sought in the simi- 
larity between the chemical and physicochemical properties of some 
tumor viruses and of certain fractions extracted from related, non- 
tumoral cells (138). This argument is in itself rather dangerous, in 
the present uncertain state of our knowledge of the structure and or- 
ganization of normal protoplasm. Any number of protoplasmic com- 
ponents could resemble viruses in gross physicochemical properties. 

Other workers have stressed more the similarity between the histo- 
logical characters of virus tumors and of normal tissues, and have com- 
pared tumor viruses ( of fowl ) to agents supposedly responsible for cell 
differentiation. Similarities between these agents and the agents that 
cause hereditary changes in bacteria have led to their being grouped 

CH. 17 Latent Viruses and Tumors 343 

together as "transmissible mutagens" (140}. The nontransmissible 
tumors would simply be produced by nontransmissible mutagens. It 
is clear that we are dealing here with the very nature of viruses and 
with their relation to the physical carriers of hereditary specificity in 
normal cells. 

Heterogenetic theories of the origin of viruses, which postulate their 
frequent rise by transformation of normal cell constituents, had been 
proposed, for example, for viruses like herpes and choriomeningitis; 
these theories were later discarded when the mode of primary infec- 
tion and transmission of these viruses was clarified. Virus hetero- 
genesis, however, is bound to remain a possibility as long as our 
knowledge of the origin of viruses and of their taxonomic position 
remains obscure. With this and related problems we shall deal in 
chapter 18. 


Origin and Nature of Viruses 

In discussing the relation of viruses to tumors we have pointed out 
the role of some viruses in determining specific processes of cell dif- 
ferentiation. This poses the problem of the relation between the 
viruses and the normal components of cells which regulate cell prop- 
erties and development, and, therefore, the problem of the natural re- 
lation and origin of viruses. 

In dealing with any group of biological objects, we must inquire 
into their relation to each other and to other objects. The justification 
for this taxonomic preoccupation is inherent in the established facts of 
biological evolution, according to which all life on earth is related by 
common ancestry. In dealing with viruses as a group, however, we 
are immediately aware that the group possesses no obvious taxonomic 
unity, but is only defined on the basis of the methodology employed 
in virus research. None of the morphological or physiological criteria 
applicable to viruses carries the taxonomic weight of the criteria used 
to group together, for example, the phylum of arthropods, or the class 
of birds, or the order of rodents. Size of virus particles, obligate para- 
sitism, and ability to penetrate some host, the criteria we used in de- 
fining viruses (see chapter 1), do not represent adequate criteria of 
natural relationship. There is no obvious evidence of a closer degree 
of relationship between any two viruses ( for example, a bacteriophage 
and a plant virus ) than between either of them and a bacterium or a 
protozoan or even any other animal and plant. There are, of course, 
close resemblances among the viruses of certain groups, such as some 
phages, the rod-shaped plant viruses, the mammalian pox viruses; these 
can be assumed to represent groups of related viruses with not too dis- 
tant common ancestors. But there the morphological, serological, and 
physiological basis of virus classification ends, and attempts to create 
meaningful virus classifications beyond this point have confused rather 
than clarified the situation. 


CH. 18 Origin and Nature of Viruses 345 

Nevertheless, it is important to locate the place of viruses in the bio- 
logical world, that is, to understand their origin and the possible paths 
of their evolution. For this, we must visualize a series of possibilities 
suggested by the known facts about viruses and other biological ob- 

Viruses as independent genetic systems. We have seen ample evi- 
dence that viruses are endowed with genetic continuity and muta- 
bility. Some viruses, moreover, may undergo a number of independent 
mutations, a fact which suggests a multiplicity of genetic determi- 
nants. Thus, viruses represent independent and specific genetic sys- 
tems. It is, therefore, unlikely that in their host cells they simply act, 
directly or through the release of inhibitions, as stimuli for the pro- 
duction of cell elements whose specificity is intrinsic in the host cell, 
or, as had been suggested, as catalysts for the maturation of already 
specific precursors (385). No such hypothesis could explain why the 
various mutant forms of a virus should cause the same cell to pro- 
duce just the corresponding type. 

We must, therefore, consider viruses as carrying more or less com- 
plex genetic systems, separate from, and in an evolutionary sense in- 
dependent of, those of the host cells which they infect from outside 
and in which they reproduce. As a matter of fact, most viruses may 
differ from their hosts genetically ( and chemically, as shown by sero- 
logical tests) as much as any two organisms taken at random may 
differ from each other. 

The chemical substrate of virus activity. What do the structure and 
composition of virus particles tell us about their nature? Ideally, a 
virus particle represents the minimum amount of material needed for 
successful invasion of a normal host. With all reservation as to the 
purity of even the best virus preparations, it is accepted that all virus 
particles contain at least nucleic acid and protein. Moreover, the 
nucleic acids and protein components of viruses contain the same 
types of constituents ( nucleotides, amino acids, etc.) as nucleic acids 
and protein of any other origin. These properties of viruses are 
unequivocal confirmation that they belong in the unique line of bio- 
logical evolution and are not new forms of life, that is, new products 
of synthesis from inert matter. The strict parasitism of viruses itself 
speaks against such an origin of viruses. Furthermore, the occurrence 
of DNA in some viruses, of RNA in others (and in some possibly of 
both) may be indicative of a relation to one or another group of cell 
nucleoproteins. It might be suggested that the nucleoprotein content 

346 Origin and Nature of Viruses 

of virus particles simply reflects the incorporation into the particles 
of parts of cell materials that are involved in virus synthesis. Nucleo- 
proteins or at least nucleic acids, however, are found wherever the 
biologist is led, by genetic or cytological evidence, to locate the ma- 
terial substrate of reproduction of specific biological elements. Since 
viruses represent systems of specificity distinct from their hosts, it is 
only reasonable to consider the nucleoproteins of virus particles as in- 
cluding the chemical materials in which virus specificity is embodied. 


Biochemical evolution. The independence and complexity of viruses 
as genetic systems lends some support to a theory of virus origin, 
widely accepted today, according to which viruses represent the re- 
sult of a regressive evolution from free-living cells. 

Studies of comparative metabolism on a variety of organisms, from 
bacteria to man, have shown that all cells share certain basic syn- 
thetic and metabolic processes and that similar intermediate metabo- 
lites are needed by all organisms as building blocks, at least at the 
nonspecific, low-molecular level (ammo acids, monosaccharides, aro- 
matic bases, coenzymes). Some organisms (autotrophs) can syn- 
thesize all their required metabolites from inorganic materials; others 
require an external supply of some organic components either as 
sources of energy ( heterotrophs ) or as building blocks (auxotrophs) 
or, in most instances, for both purposes. It has been suggested (504) 
that the earliest self-reproducing units, the primitive life forms, came 
into existence in an environment rich in organic compounds, which 
were produced by nonenzymatic synthesis in the cooling earth utilizing 
the energy of ultraviolet light. Presumably, the primitive forms of 
life were dependent for their reproduction on the external supply of 
organic building blocks. They could have acquired more and more 
synthetic power by mutation (343), and the synthetically more e- 
endowed forms could have been selectively favored as the external 
supply of organic substances dwindled, thus leading to the establish- 
ment of autotrophs. 

In the course of successive evolution, some of this synthetic power 
was often lost again and in today's world most animals and micro- 
organisms, and many plants, are to some extent auxotrophic and het- 
erotrophic. To this "regressive evolution" (450) has been attributed 
a more prominent role in biochemical evolution than it probably 
deserves except in the restriction of most organisms to certain eco- 

CH. 18 The Regressive Theory of Virus Origin 347 

logical environments. Biochemical differentiation must have depended 
mostly on the appearance of new patterns of synthesis, producing new 
specific macromolecules (proteins, polysaccharides, nucleic acids), 
the chemistry and metabolism of which are still obscure. 

Viruses as regressed parasites. A case has been made ( 114; 271; 395) 
for an origin of viruses, or at least of some viruses, by an extreme 
process of regressive evolution. Free-living microorganisms that re- 
quire a number of preformed metabolites would become parasitic on 
other organisms that act as suppliers of the required substances. Para- 
sitism, and especially intracellular parasitism, would favor further loss 
of synthetic abilities by providing an environment where most needed 
metabolites are present, thus placing little or no handicap on the syn- 
thetically more deficient mutant forms. It is argued that this process 
could proceed to such an extent as to cause the loss of much or most 
of the enzymatic machinery of the parasite, with accompanying reduc- 
tion in size and transformation into a virus. A virus could ultimately 
be reduced to the naked genetic system (or possibly even a part of 
it) of a formerly free-living organism. 

Appealing as this theory may be (especially to the frequent be- 
lievers in an intrinsic, mystic tendency to decay in nature) there is 
no strong evidence to support it. Although there is some parallelism 
between the degree of parasitism and the nutritional requirements of 
facultative parasites (bacteria, fungi, and protozoa), hardly any tran- 
sitional stages are known between viruses and obligately parasitic 
protozoa such as the malarial plasmodia, or between viruses and obli- 
gately parasitic bacteria such as the leprosy bacillus. Rickettsiae (see 
chapter 19 ) are small intracellular parasites, probably requiring a large 
number of specific supplies from their host cells; but their relationship 
to any viruses is doubtful. 

In many arthropods are found a variety of intracellular symbionts, 
some yeastlike, some bacterialike (bacteroids, 62(5), which are trans- 
mitted from mother to young by infection of the eggs and which seem 
to perform some functions needed by the hosts. In spite of their long 
association with their hosts (shown by the similarity of the symbionts 
in insects that have long since branched off from common ancestors, 
such as roaches and termites), the symbionts, whose function is un- 
known but probably essential to the life of the hosts, have hardly re- 
gressed at all as far as size and morphological complexity are con- 
cerned. The same is true of the algae (ZoochloreVa) symbiotic with 
certain protozoa. 

348 Origin and Nature of Viruses 

It is hard to see how regressive evolution through parasitism could 
lead to the establishment of such an intimate organizational relation 
between parasite and host as that found in bacteriophages and prob- 
ably in many other viruses, a relation in which the virus seems to 
integrate itself into the synthetic and genetic machinery of the host 
and to determine in some cases the specificity of the whole product 
of synthesis. 

The particles of some animal viruses, such as vaccinia, in the most 
purified preparations contain substances like riboflavine and biotin, 
known to play coenzyme roles in metabolic systems present in all 
cells (606). This has been taken to suggest that the virus is derived 
from a free-living form possessing its own metabolic machinery, and 
that the coenzymes represent remnants of this machinery. We cannot 
exclude, however, the possibility that these substances derive from 
the host cell and cannot be separated from the virus particles, because 
of the way the mature virus particle is formed. The mature virus 
particle, defined as the minimum piece of material needed for infec- 
tivity, may well incorporate some host components along with the 
essential, specific virus substance (142; 377). 

For viruses whose particles contain RNA rather than DNA, the 
hypothesis of regressive evolution leads to the idea that the virus 
represents a remnant of some extrachromosomal element of the free- 
living ancestor*: The fact that all the plant viruses thus far analyzed 
contain RNA favors, instead, the idea that these viruses originated by 
some common mechanism inherent in the plant cell possibly by trans- 
formation of plastids. 

There is, of course, no reason to believe that all groups of viruses 
originated by the same process; regressive evolution might account for 
the origin of some of the largest and most complex animal viruses. 
Also, we should bear in mind that the above considerations only apply 
insofar as the regressive evolution theory is taken to imply a relatively 
recent origin of viruses from free-living forms similar to those known 
to exist now. The possibility that in the course of evolution free- 
living forms have merged with cells more or less distantly related and 
can again emerge and become viruses, can better be discussed after 
we consider the possible origin of viruses from cell components. 


Viruses and cell constituents. Many workers have emphasized the 
similarity of viruses to some cell components and have inclined to the 

CH. 18 The Relation of Viruses to Cell Constituents 349 

belief that the viruses have derived from cell constituents that have 
become transmissible from cell to cell that is, infectious. Sojne of 
the supporting evidence, as far as virus research is concerned, has 
been discussed in chapter 17. We may recall the examples of virus 
latency and masking, the morphogenetic effects of some tumor viruses 
in stimulating cell differentiation, and the similarity between the tu- 
mors caused by viruses and those from which no virus can be ex- 
tracted. The advocates of the origin of viruses from cell constituents 
stress the occurrence of diseases of plants, supposedly caused by 
viruses but transmissible only by graft, as evidence for the direct 
transmission from cell to cell of "infectious cell proteins." Opponents 
of the theory have objected, first, that several virus diseases sup- 
posedly transmissible only by graft have finally yielded bona fide 
viruses (52). Second, in most cases of virus latency in animals or 
plants some virus can be shown either to have entered the organism 
very early in life, or to have been transmitted through egg, sperm, 
or pollen. 

The insect virus that causes silkworm jaundice, a polyhedral dis- 
ease, can manifest itself in normal silkworms following feeding with 
various chemicals ( hydroxylamine, acetoxime, potassium nitrite; 690a). 
The claim has been made that the virus is produced by a transforma- 
tion of the genes of the host into virus. It seems more reasonable to 
assume that the chemicals employed can induce the maturation and 
pathogenic action of a latent virus (provirus) transmitted from gen- 
eration to generation of silkworms through the egg (654a). 

The question, however, is deeper. If a virus can remain latent in 
the cells of a host for many cell generations and even through the 
meiotic process, it must be integrated in the normally functioning cell 
machinery and as such it is a cell component. The question should 
be put in a different way: Have all viruses, as we know them, entered 
more or less recently (from one to several thousand cell generations) 
the lineage of host cells in which we find them, or have they a direct 
genetic relation with the materials proper to some host cells? More- 
over, are these two alternatives mutually exclusive? That is, can the 
distinction between "genetically specific components of a cell" and 
"exogenous specific self -reproducing elements" be legitimately Ynade? 
It is clear that a meaningful approach to this problem cannot be made 
from the viewpoint of virology alone, but must be based on a knowl- 
edge of the genetic organization and evolution of the cell. 

Viruses and chromosomal genes. As far as we are aware, all cells 
derive from cells. Although the cell is the only unit of reproduction 

350 Origin and Nature of Viruses 

observable in isolation, we trace the seat of the specificity of a cell to 
certain materials it contains. Mother and daughter cells present both 
similarities and differences, the differences manifested in the phe- 
nomena that go under the general headings of development and dif- 
ferentiation. In discussing cellular organization for our purpose, we 
should remember that we are looking for something that may be re- 
lated to viruses. Since viruses can exist as discrete material particles 
carrying more or less complex systems of hereditary determinants, it 
is only natural to look in the cell for something having properties of 
this type. 

The first cellular elements to be considered are the genes, located 
in the nuclear chromosomes. The genes are determinants of heredity, 
transmitted in an orderly manner from cell to cell at mitosis. Their 
linear arrangement in the chromosome strings is proved by cyto- 
genetical observations on crossing over, chromosomal breaks, and re- 
arrangements of various kinds (604). As a rule, each gene, identified 
by the effects of its changes on the properties of the organism, is 
located at one specific position in a chromosome. The chromosomal 
location of DNA and other suggestive evidence for an actual role of 
desoxyribonucleoproteins in the hereditary mechanism of chromo- 
somes emphasize an important similarity between genie material and 
viruses such as bacteriophages and most animal viruses. Actually, the 
complex recontbination phenomena observed with bacteriophages ( see 
chapter 9) and more recently with influenza viruses (120) closely re- 
semble the behavior of gene groups such as we find in chromosomes. 
The multiple mutability of many other animal viruses suggests that 
they too are genetically complex. These similarities between viruses 
and chromosomal genes may indicate a recent common origin, and the 
idea of viruses as "naked genes" was suggested as early as 1920 (192; 
496; 683). It would now seem more fitting to compare some viruses 
with naked gene complexes, chromosomes, or even groups of chromo- 
somes or nuclei. 

A rudimentary nuclear system could have derived either by regres- 
sive evolution through parasitism or by acquisition of transmissibility 
froifc cell to cell by a whole or partial nucleus. The first possibility 
has been discussed above; is there any evidence for the second? The 
most suggestive evidence is provided by a variety of transformation 
phenomena that occur in bacteria. 

CH. 18 Viruses and Genetic Transfers in Bacteria 351 


Viruses and transforming principles. Type transformation is classi- 
cally observed in pneumococci (see 32). Diplococcus pneumoniae oc- 
curs in a multiplicity of smooth types ( SI, SII, SIH, ) each char- 
acterized by the presence, amount, and chemical nature of a capsular 
polysaccharide. A stable noncapsulated (R) form can be derived 
from any S type by mutation. The R strain can be transformed into 
a hereditarily stable capsulated form, producing a specific polysac- 
charide, by growth in the presence of an extract of a given capsulated 
type. For example, an R strain derived from SII can be transformed 
into a strain SIII by growth with an extract of SIII cells. 

The change from R to S is not the result of a selection of sponta- 
neous bacterial mutants. It is actually induced by the extracts in an 
appreciable proportion of the exposed cells. The responsible agents or 
transforming principles are present in a DNA fraction of the extracts. 
The principles are not separable chemically from the bulk of the 
DNA. Presumably they represent, in part or full, the genetic deter- 
minants responsible for the production of the capsular material. The 
transformed organisms become, in turn, a permanent source of the 
corresponding transforming principles. Other characters, such as sen- 
sitivity or resistance to penicillin, can also be transferred by means of 
transforming principles in the DNA fraction (350a). 

Recent developments (210) indicate additional similarities be- 
tween transforming principles and chromosomal genes or gene com- 
plexes. The principles from different serological types (SI, SII, 
SIII, ) are as a rule mutually exclusive, just like allelic forms of a 
certain gene. Nevertheless, interactions occur between slightly dif- 
ferent forms of the same transforming principle, with results com- 
parable to recombination phenomena. For example, two different 
principles acting together can induce into an organism a third, novel 
type of property, which is afterward transmissible hereditarily (or by 
transforming principle) without segregation into the two parental 

Here and in similar phenomena with Hernophttus influenzae (11), 
we have the closest thing to the exogenous transmission of hereditary 
determinants of the type usually associated with genes. Notice that 
the transforming principles fulfill our definition of viruses (or, in the 
cell, of proviruses ) . Yet, only a fiend for semantics would consider the 

352 Origin and Nature of Viruses 

capsular types in pneumococci as induced by viruses, at least as long 
as the word virus is used in the commonly employed sense. We are 
clearly at the borderline. 

Bacteriophages and bacterial genes. Recent work has established 
even closer ties between the genetic determinants of bacteria and the 
viruses of the bacteriophage group. In the first place, we should re- 
call the increasing evidence in favor of a chromosomal localization 
and transmission of the prophages of lysogenic bacteria ( see page 205 ) . 
The determinant elements of these latent viruses may find their way 
into the nuclear apparatus of their host and integrate themselves 
within it. 

In the second place, recent observations implicate bacteriophages in 
the transmission of genetic properties from one strain of bacteria to 

Transduction in Salmonella. Some bacteria of the genus Salmonella 
are lysogenic and liberate in their culture media a filtrable agent iden- 
tical with the phage and which can transfer to other strains some 
hereditary characters of the strain from which it comes (695). The 
transferred characters include fermentative properties, antigens, and 
sensitivity or resistance to chemical agents. This transduction of 
properties ordinarily involves only a single character at any one time, 
as though single determinants of heredity were transduced from strain 
to strain. 

We may assume that in the process of phage maturation some phage 
particles can receive not only their own prophage but also a portion 
of a chromosome. Immediately, we are led to wonder what the "own" 
prophage is. Could any part of the host chromosomes be a potential 
genetic core for a transmitting agent? Such an agent might only mani- 
fest itself as a phage if it happened to be endowed with lytic prop- 
erties; otherwise, it would appear as a "pure" agent of genetic transfer. 
The transducing agent of Salmonella might indeed act on some strains 
as a pure gene-transferring entity, while acting as prophage carrier on 
the phage-sensitive cultures. 

The fertility agent of Escherichia coli. Genetic recombination in 
E. coli K-12 has recently been shown to be a nonsymmetrical phe- 
nomenon, which requires that one or both strains possess a certain 
fertility factor (F+ strains). This factor can be transmitted infec- 
tiously from F+ to F~ strains by growth in mixed culture. The fer- 
tility factor is not filtrable; it may be extruded onto the surface of the 
F+ cells, from which the F~ cells presumably can pick it up and be 

CH. 18 Viruses and Genetic Transfers in Bacteria 353 

transformed by it into F+ cells (295; 413). Thereafter, the fertility 
factor continues to be produced in the transformed cells. 

Bacteriophage infection and genetic transfers. Bacteriophage re- 
search has shown that phages contain a genetic portion, which becomes 
infectious at maturation by acquiring a protective skin. The skin in 
turn mediates the introduction of the genetic portion into another host 

It seems justifiable to suggest that there may be a whole series of 
cases whose common feature is the transfer of genetic characters. 
The vehicles of such transfer may or may not carry lytic principles 
recognizable as phages. At one end we may have the transforming 
principles of pneumococcus and other bacteria, representing genetic 
determinants that can withstand artificial extraction and can enter as 
such into recipient cells. Transduction in Salmonella may require the 
inclusion of the genetic material to be transferred into carrier elements 
similar to those acquired by phages at maturation. In different in- 
stances the transferred genetic material may perform differently in the 
recipient cells. It may take over the nuclear controls and disrupt the 
cell (lysis by virulent phages), or it may introduce a latent tendency 
to ultimate disruption (lysogenesis for temperate phages), or it may 
introduce one or more properties derived from the cell in which it 
was formed. Lytic (incompatible) and hereditary (compatible) de- 
terminants may be included together in a single carrier. Ultimately, 
we may encounter cases in which the whole genome of the parent cell 
will be introduced into the host cell, and the carrier element will be- 
come analogous to a sperm cell. 

At the other extreme, the carrier may become a purely bactericidal 
agent, such as a colicine (see page 207), which resembles phages but 
fails to reproduce in the host it destroys. 

Not only genetically, but also biochemically, there may exist a series 
of agents intermediate between metabolically inert virus particles and 
metabolically active sperm cells. Indeed, the L forms of bacteria 
(including elements at the limit of microscopic visibility, copiously 
formed by bacteria under special circumstances; J79) may represent 
such intermediate forms and may be involved in some form of trans- 
duction of genetic characters. 

Thus, the phages appear to be close relatives of the genetic material 
of the host cells; so close indeed that it becomes hard to distinguish 
between a prophage and a portion of a bacterial chromosome, or be- 
tween a virulent phage and an incompatible portion of genetic mate- 
rial. The "parasitism at the genetic level" of virulent phages (437) 

354 Origin and Nature of Viruses 

would only be one extreme in a series of possible levels of genetic 
integration between virus and host. 

Should we consider the phages as reproductive organs of bacteria, 
and in a different category from all other viruses? Or do other trans- 
missible agents defined as viruses have the same type of relationship 
to their host as phage? No factual evidence for a role of plant or 
animal viruses in the hereditary mechanisms of their host cells is avail- 
able. Yet, the latency of many viruses, their deep integration in the 
host-cell machinery, and their directing role* on cell properties, espe- 
cially clear in tumor viruses, suggest that the possibility of a relation 
between viruses and host-cell heredity should seriously be entertained. 
In an attempt to find possible relatives of these viruses, let us return 
to our search for viruslike elements in normal cells. 


Outside the nucleus, the cell contains a cytoplasm, whose fine 
structure is largely unknown, but which contains a variety of recog- 
nizable, differentiated, discrete elements, some microscopic (centro- 
somes, mitochondria, plastids), others inframicroscopic (microsomes 
(178). Some of the cytoplasmic elements reproduce and are dis- 
tributed at cell division to the daughter cells, either with some regu- 
larity, so that v gross inequalities are avoided ( as with the mitochon- 
dria), or with great regularity, as in the case of the centrioles, which, 
when present, duplicate and segregate as a part of the mitotic figure. 
For comparison with viruses, however, we need something more than 
cytological evidence of granular nature and ability to divide. Ac- 
tually, a granular structure is not a necessary element of similarity to 
viruses, since inside the host cell the viruses may not be present in 
any recognizably granular form. What is needed is evidence for cyto- 
plasmic elements endowed with individuality and intrinsic genetic 
continuity. Individuality means that one element is sufficient to de- 
termine the production of copies of itself, just as a virus or a gene. 
Intrinsic genetic continuity means (0), that the element is an indis- 
pensable source for the production of more elements like itself; and 
( b ) , that the specific configuration of the new elements is determined 
by the original element itself. The reason for making the criteria so 
specific is that viruses fulfill these criteria, whereas among the ele- 
ments involved in the so-called cytoplasmic inheritance we find listed 
entities of different types, which may or may not meet our require- 

CH. 18 Viruses and Cytoplasmic Inheritance 355 

Cytoplasmic inheritance (126) is defined as the transmission from 
generation to generation of characteristics determined by the cyto- 
plasm rather than the nucleus. It manifests itself in sexual organisms 
as a difference in the result of reciprocal crosses, in which the nuclear 
complements are similar but the cytoplasms, mostly maternal, are dif- 
ferent. The new phenotypes, produced by the same genes working 
in different cytoplasms, show some of the characters of the maternal 
parent. Cell differentiation, in which cells with presumably identical 
nuclei develop into a variety of types, has also been attributed (hypo- 
thetically) to Cytoplasmic influences. 

In itself, Cytoplasmic inheritance does not prove that the cytoplasm 
contains hereditary determinants responsible for the reproduction of 
their own specific materials. The cytoplasm is a complex and prob- 
ably highly organized mixture of organic substances, always reacting 
with one another, some being destroyed, others formed. The nature 
and rate of production of various substances may be determined 
mainly by the nucleus, considered as the bearer of the specific patterns 
for the structure of the proteins and other complex organic molecules 
whose production it determines. This is exemplified by the determi- 
nation of enzyme and antigen specificity by chromosomal genes (687). 
The gene products will, of course, be affected by the Cytoplasmic 

The mechanism of Cytoplasmic inheritance could be a "molar" 
rather than a "molecular" one, depending not on specific individual 
determinants but on the relative amounts and distribution of inter- 
acting substances in the cytoplasm. The cytoplasm is conceivably 
capable of reacting in a variety of ways to chemical stimuli derived 
from the nucleus, according to its momentary status (171). Specific 
cytoplasmic reactions may thus appear without the intervention of 
specific cytoplasmic determinants of heredity. 

The cytoplasm contains characteristic granules, which divide regu- 
larly at or before cell division. These include the centrioles, the 
mitochondria, and the plastids, and the kinetosomes of ciliate protozoa 
(that is, the granules from which cilia and trichocysts originate) 
(451). Such visible, regularly dividing granules are not necessarily 
the carriers of discrete specific determinants; they could be function- 
ally individualized centers for the accumulation of materials of nuclear 
origin, whose intrinsic specificity could be gene-determined rather 
than cytoplasmically determined. 

Sonaeborn has pointed out (617) that the essential requirement for 
the recognition of a cytoplasmic element as a specific self-reproducing 

356 Origin and Nature of Viruses 

genetic determinant is the proof of mutability and reproduction in the 
mutated form, independent of environmental or genie uniformity. 
Only the ability to reproduce a new pattern of specificity, acquired 
by mutation, in the presence of any compatible genie background, 
represents adequate proof for the intrinsic specificity of a cytoplasmic 
determinant of heredity. Some such determinants are known to exist, 
and we shall now discuss them briefly. 

Plasmagenes in relation to viruses. In plants, the plastids, cyto- 
plasmic bodies containing chlorophyll in the green parts of the plants, 
are transmitted through the egg-cell protoplasm and occasionally 
through the pollen. They are mutable; green plastids occasionally give 
rise to colorless ones, and the resulting variegation is self -perpetuating 
( 551 ) . Nuclear genes control both the mutability and the reproduc- 
tion of plastids, but not their quality. A gene may increase the rate 
of plastid mutation or make it impossible for a mutated plastid to 
reproduce, but there is no gene pair responsible for the regular pro- 
duction of two alternative forms of plastids. The relationship of 
plastids to mitochondria is generally admitted; but the proof of the 
genetic continuity of mitochondria would require the discovery of 
intrinsic mitochondrial mutability. 

Some races of the protozoon Paramecium aurelia contain granules 
(0.2-0.8 fi) that give the cytochemical reactions of DNA and are iden- 
tified with the \appa factor. This factor regulates both the secretion 
of a poison (paramecin, a substance that contains DNA and is toxic 
for individuals of other races ) and the sensitivity of the animal to the 
poison (616). Animals with the kappa factor produce the poison and 
are resistant to it. In the proper genetic background (presence of a 
dominant gene K and of accessory genetic influences) an animal con- 
taining the kappa factor will produce more of it, an animal without 
kappa (such as can be obtained, for example, by irradiation) will not 
produce any: thus, kappa is self -reproducing. By proper cultural 
conditions, the growth rate of the animal and of kappa can be dis- 
sociated, and kappa may then be totally and irreversibly lost in some 
offspring. But as long as one unit of kappa remains, the factor can 
be restored to its normal amount of several hundred or several thou- 
sand units: thus, kappa is genetically individualized. Moreover, kappa 
is mutable. A mutated kappa will cause production of a different 
paramecin, recognizable by its toxic effects, and the animal with the 
mutated kappa will be specifically resistant to the corresponding para- 
mecin. No gene difference is involved. Kappa can be introduced into 
suitable paramecia by contact with extract of kappa-bearing animals, 

CH. 18 Viruses and Cytoplasmic Inheritance 357 

and it reproduces within the new host. Thus, kappa is potentially 
infectious and represents another borderline case with viruses. 

A similar situation exists in the case of cytoplasmic inheritance of 
CO 2 sensitivity in Drosophila (421). Sensitive flies are killed or in- 
jured by a short exposure to CO 2 , whereas resistant flies are not 
affected. The agent, sigma, is self-reproducing, individualized, and 
mutable. Sigma is regularly transmitted through the egg and oc- 
casionally through the sperm, and is also transmissible by cell-free 
extracts of organs of CO 2 -sensitive flies. Sigma has many properties 
of a virus. Its cycle of reproduction in flies infected by injection is 
similar to that of viruses. The interactions between sigma and its 
mutants resemble those between a temperate phage and its virulent 
mutants. Sigma is inactivated by doses of x-rays comparable to those 
required for inactivation of medium-sized viruses (422), whereas 
kappa is as sensitive to x-rays as a bacterium (529). This difference 
may not be significant, however, since kappa was irradiated intra- 
cellularly, whereas sigma was irradiated in free state. Ultrafiltration 
data indicate that sigma consists of particles larger than was estimated 
from x-ray experiments (421). 

These examples indicate beyond doubt the existence of relatively 
independent hereditary determinants in the cytoplasm of certain or- 
ganisms. The term plasmagene (617) is often used to designate such 
determinants. It has been extended to cover other cases of cytoplasmic 
inheritance, in which the existence of such determinants is suspected 
but not proved. 

For example, in Paramecium aurelia itself, cytoplasmic inheritance 
is present for a variety of specific, mutually exclusive antigenic types, 
recognizable by the immobilization of animals of a given type by type- 
specific antiserum (617). Exposure to homologous antiserum (or 
other treatments ) can transform the cells into another serotype, within 
the spectrum of antigenic potentialities of the race. Here, however 
(at variance with the situation for the killer factor kappa), the 
specificity of the alternative serotypes seems to be determined by 
nuclear genes. The cytoplasmic effect is due to the fact that whatever 
serotype happens to be present in a given clone will persist unless 
changes occur either in the genetic make-up (by crossing) or in the 
environment (for example, changes in temperature or contact with 
specific antisera ) . 

In yeast, a mutation that brings about the loss of aerobic respiration 
and a change in colony size can occur either spontaneously or upon 
treatment with acriflavine (209). The mutant cells lack certain recog- 

358 Origin and Nature of Viruses 

nizable cytoplasmic particles analogous to mitochondria and containing 
the missing respiratory enzymes (cytochrome oxidase and succinic 
dehydrogenase ) . The nuclear genes are unchanged by the mutation, 
as proved by crosses, and have not lost their ability to support the for- 
mation of the missing enzymes. Here again, cytoplasmic inheritance 
is present, but the existence of specific determinants in the cytoplasm 
is unproved. The cytoplasmic elements may receive their specificity 
from the nucleus, and their role in controlling the production of more 
elements like themselves may be a stabilizing rather than a determining 

What we have said can be summarized in the statement that the 
cytoplasm is often involved in the determination of specificity, but 
only in a few cases has been proved to contain discrete genetic 
determinants responsible for their own intrinsic specificity. Most other 
instances of cytoplasmic inheritance could be due to self-maintenance 
of alternative states of the cytoplasm through interactions among non- 
self-reproducing elements. 

Adaptive enzymes, as analyzed in bacteria and yeasts, are instructive 
in this connection (487; 488; 619). The production of an adaptive 
enzyme is made possible by the presence of a specific gene and by 
the presence of the substrate or of a variety of substances (inducers) 
related to the substrate. The rate of formation of the enzyme is of 
an exponential, apparently autocatalytic type, increasing as the amount 
of enzyme already present increases. It is tempting to suggest that 
the enzyme, or a precursor of it, is self-reproducing. However, an 
exponential rate of synthesis is not necessarily a sign of self-reproduc- 
tion. For example, all cell proteins increase exponentially in bacterial 
cells during the phase of exponential cell growth. This is only an 
expression of the overall increase in protoplasm. An adaptive enzyme, 
formed in growing cells, will share the growth rate of the protoplasm 
as a whole. In nongrowing cells the initial formation of an enzyme 
may release an inhibition on its further synthesis. 

A true self-determination of an enzyme system seems to have been 
demonstrated in yeast (620; 621 ). Enzymatic adaptation to ferment a 
given sugar requires the presence of substrate and of a given gene; but, 
provided the substrate is present, enzyme activity persists for many 
generations after the gene is removed by crossing to a nonadapting 
form. The main interest of this observation is the suggestion it pro- 
vides of a self-determination of gene-derived elements, that is, of the 
possible existence of "gene-initiated plasmagenes." 

CH. 18 Virus Origin and Cell Origin 359 

Returning to the well-established cases of mutable plasmagenes 
(kappa, sigma, plastids) it is clear that they share with viruses some 
essential properties, such as intrinsic mutability and dependence on 
specific host genes for reproduction, but not for initiation. Gene- 
directed mutability in viruses has not yet been demonstrated, but this 
may be due to inadequate study. The formal analogy goes beyond 
this. Both kappa in Paramecium and sigma in Drosophila can be 
transmitted from one organism to another by cell-free extracts. 
Actually, our definition of virus is fulfilled; kappa ( apart from its large 
size, similar to that of rickettsiae) and sigma are as good "viruses" as 
a phage or a potato virus. The transmission of kappa or sigma by man- 
made extracts to properly receptive animals is, of course, an "un- 
natural" event. But is it more "unnatural" than the liberation of 
phage by a lysogenic bacterium and its chance encounter with a 
susceptible cell, or than the grafting of the King Edward potato, 
carrying the paracrinkle virus, onto a potato stock susceptible to the 

In guinea pig, black pigmentation of the skin is due in part to the 
presence of ramified cells carrying melanic pigment ( melanophores ) . 
Upon transplantation of black skin into a white skin area, black pig- 
mentation spreads to the white area, apparently by actual transforma- 
tion of colorless melanophores into pigmented- ones. This transfor- 
mation can be propagated indefinitely by successive grafts ( 78 ) . The 
unknown agent is transmissible by graft, as many plant viruses are, 
although not yet so by cell-free extracts. There is not much sense, of 
course, in calling the pigment factor in guinea pig skin a virus. Be- 
cause of the well-known and orthodox Mendelian inheritance of skin 
color in guinea pig, the nature of the pigment factor as a cell com- 
ponent, and its gene-derived specificity, are not easily questionable. 


It is at the point where restrictive definitions break down that we 
must search for meaningful tie-ups between viruses and other self- 
reproducing agents such as genes or plasmagenes. The matter of 
definition has been dismissed as immaterial by some authors concerned 
with cytoplasmic inheritance, but is very important for the virologist. 
The crucial question for the student of plasmagenes is that of repro- 
ductive independence; for the virologist, however, it is that of taxo- 
nomic independence. A virus may enter a host cell and remain in it 
for one cell generation or less (virulent phage), or it may enter and 

360 Origin and Nature of Viruses 

remain through many cell generations (prophage in lysogenic bac- 
teria; tumor viruses; plant viruses; etc.). In sexually reproducing 
organisms, a recently entered virus may be transmitted through the 
egg (or the sperm or pollen grain), and the next generation may or 
may not show signs of virus presence. Thus, virus latency is a matter 
of duration; a long-latency virus is practically indistinguishable from 
a cell component. It could be called a virus or a plasmagene ( or even 
a gene), according to the effects by which it happened to be detected, 
that is, an apparently "normal" manifestation or a clearly abnormal one. 
With lysogenic bacteria, a virus-infected cell may acquire character- 
istics that could also be acquired by genetic mutation, as in transduc- 
tion phenomena in Salmonella. 

Concerning virus origin, the problem is, then, 2-fold. It involves, 
(a) the relation of viruses to cell components, and (b) the origin of 
cell components. 

The gene complement of a cell (its genome) is often supposed to 
have originated by differentiation of an original self -reproducing single 
element, copies of which have accidentally remained together and by 
mutation have assumed different forms and functions. These gene 
groups would have progressively developed into chromosomes, be- 
cause of strong advantages that an organized mechanism for equiparti- 
tion of the genetic material offers in preserving favorable gene com- 
binations. Sex\ial phenomena, arising later in evolution, would bring 
about further complications, but in presexual organisms all genes would 
have originated within the single cell lineage (monopkyletic). Several 
cases are actually known of genes with similar function, possibly 
derived from one another and situated together in the chromosome 
(420). The cytoplasm, in the simplest hypothesis, would be derived 
from the activity of the genes. As a result, any genetically specific cell 
component should be a close relative of all others within one genetic 
line. The transmission of such a component gene, chromosome, or 
plasmagene to another cell would represent a merging of a portion of 
a genetic line into another line, a merging that might lead to para- 
sitism ( virus formation ) or to symbiosis ( plasmagene ) . 

On the other hand, the normal cell might be polyphyletic in origin. 
It is conceivable that several primitive self-reproducing molecules may 
have come together into a successful combination and developed into 
a cell; or that some such elements may have entered an already formed 
cell. The plastids may well be such late comers, not yet integrated 
into the chromosomal mechanism. The merging of genetic lines would 
then have occurred early in the evolution of the cell, and the fact of 

CH. 18 Virus Origin and Cell Origin 361 

a gene, chromosome, or plasmagene later becoming transmissible from 
cell to cell might represent a reacquisition of the original independence 
and a repetition of the original merging process. 

The "viroid" theory (12) specifically postulates such an entry of 
free, primordial, self-reproducing elements into cells early in evolution, 
the preservation on their part of primordial characteristics, and their 
ability to regain infectivity by mutation, thus giving rise to viruses. 
Mutations of viroids could also give rise to nontransmissible, abnormal 
plasmagenes and be responsible, for example, for the tumoral trans- 
formation of cells. The theory is somewhat more restrictive than 
another conceivable one, which would include, among the potentially 
transmissible cell constituents, the nuclear genes as well as the cyto- 
plasmic components. 

Thus, all speculation as to virus origin leads to the possible modes 
of merging of two genetic systems into a functioning cell. The merg- 
ing is not obvious in the case of rapidly destructive viruses. Its funda- 
mental importance has escaped some virologists, accustomed to think- 
ing of viruses in a cell as of bacteria in a culture. Yet, even a bac- 
terium infected with a virulent phage and doomed to early lysis is a 
functional, integrated cell system, whose ultimate fate lysis and dis- 
integrationis incidental to the primary event of genetic and bio- 
chemical integration of viral and cellular machinery. 

Merging leads to a more lasting integration with viruses that are car- 
ried for several cell generations or even through the sexual process. 
It may become permanent, as in virus masking, for example, in the 
tumors that papilloma virus induces in the domestic rabbit (602), or 
with egg-transmitted latent viruses such as lymphocytic choriomenin- 
gitis ( 651 ) . Some plasmagenes may have originated in this way. On 
the other hand, the facts of lysogenicity and their relation to genetic 
transfers in bacteria suggest possible paths of the evolution of sexual 
mechanisms from viruslike mechanisms ( and vice versa ) . 

Ultimately, the taxonomic position of a certain element will depend 
on the duration of its joint evolutionary history with other components 
of the same cells. The ability to regain independence may be a func- 
tion not only of intrinsic mutability but also of the duration of the 
partnership, which may force an ever-increasing interdependence 
among the various cell components. There is the danger, in such an 
expression as the "viroid" theory, of assuming a lasting primitive quality 
in the "viroid" component of the system as compared with the evolving 
quality of the rest of the genetic materials of the cell. An exogenous 

362 Origin and Nature of Viruses 

element, placed within a cell lineage, would probably undergo as much 
evolutionary change as the other genetic elements of the cell and 
should not resemble its primitive ancestor any more than the other 
elements resemble theirs. 

The basic similarity of viruses to other biological elements has thus 
led us to a somewhat unified viewpoint, by showing us that different 
theories of virus origin differ only in their interpretation of the relative 
duration of the companionship between cell components. Merging 
of genetic lines and integration at the cellular level are the common 
denominator of all theories. A virus could derive from any one (or 
several) of the genetically specific components of a cell, either by 
regressive evolution or by transmission of a portion of a cell to another 
cell (by contact between intact cells or between cells and cell extracts). 

If the transmitted element proved to be rapidly destructive for the 
new cell system it would generally be lost, and could only maintain 
itself in existence through the availability of innumerable suitable hosts. 
Often, however, the merging could last. The meaningful question 
remains and cannot be answered at present whether such a merging 
is a novel and exceptional feature, leading mainly to the formation of 
abnormal complexes of low evolutionary value (diseased organisms) 
or if it is an example of a process that has played and is still playing an 
important role in Devolution and possibly in development. 

In truth, a virus may be both a regressed parasite and a cell com- 
ponent that has become infectious, depending simply on which phase 
of the evolutionary history of its genetic material we are observing. 
It may have been both things at different times. 

The intrinsic properties of certain viruses may suggest a closer rela- 
tion to one or another type of genetic determinant. Thus, bacterio- 
phages, with their complex genetic apparatus, may be closer to systems 
that have already evolved into chromosomal complexity and may 
actually be manifestations of evolving sexual mechanisms. Plant 
viruses, which contain RNA, may be more closely related to plastids 
or, if any such exist, to other RNA-containing plasmagenes. Animal 
viruses containing DNA may be akin to nuclear determinants. 

In conclusion, we see that, just as the study of virus structure and 
multiplication always leads us back to the cell as the system in which 
the phenomena of life take place, so does the problem of virus origin 
lead us back to the origin of the cell as an integrated whole. A 'virus 
is nothing but a part of the cell. We observe and recognize as viruses 
those parts independent enough to pass from cell to cell, and we com- 

CH. 18 Virus Origin and Cell Origin 363 

pare them with other parts that are more tightly tied up with the whole 
system. It is indeed this aspect of viruses that makes them invaluable 
to the biologist, whom they present with the unique opportunity of 
observing in isolation the active determinants of biological specificity, 
which are truly the stuff of which all life is made. 



The Rickettsiae 

A group of microorganisms, the rickettsiae, is often described to- 
gether with viruses. This is because its members share some of the 
properties by which viruses are characterized, namely, the small size, 
the obligate intracellular parasitism, and the ability to produce ab- 
normal manifestations in certain hosts. The main similarity between 
viruses and rickettsiae is their obligate intracellular reproduction or, 
at least, their inability to reproduce in any of the cell-free culture media 
that have been tested. The similarity between the rickettsiae and 
viruses does not go very far, however. Because of their larger size and 
microscopic visibility, the rickettsiae have raised few of the methodo- 
logical and operational problems involved in defining the field of 
virology. They resemble bacteria much more than they resemble 
viruses, and there is no definite reason to suspect that they represent 
a direct evolutionary link between the two groups. At the same time, 
they are an extremely interesting biological group, because their 
parasitism in certain hosts and their symbiotic relation with others 
throw some light on the evolution of parasitism and on the epidemio- 
logical evolution of microbial diseases (34). 

Most efforts to classify the rickettsiae have been limited to those 
recognized as causing disease in man or in other vertebrates (518). 
This choice, however, is largely artificial, because the pathogenic 
rickettsiae are only a small sample from a large group of microorgan- 
isms symbiotic with a variety of arthropods (insects or arachnids) 
(626). Only a few of these microorganisms accidentally embark upon 
a pathogenic career by being transmitted from the arthropods into some 
other animals, in which they may cause disease. It is, therefore, prob- 
able that some of the rickettsiae found in diseased hosts are only dis- 
tantly related among themselves and that their close relatives should 


CH. 19 

Structure of Rickettsiae 


be sought among the symbiotic inhabitants of arthropods. Only in a 
few instances, such as for epidemic and murine (endemic) typhus, is 
there reason to think that the causative agents have evolved from one 
another since their establishment in vertebrate hosts. 

The practical importance of rickettsiae is that they give rise to several 
important diseases, among them epidemic and endemic typhus, Rocky 
Mountain spotted fever, and tsutsugamushi disease (scrub typhus). 
The rickettsiae pathogenic for man are listed in table 32. 

Table 32. Rickettsiae pathogenic for man (family Rickettsiaceae) 

Slightly modified from Pinkerton (51 8) 


Species Variety 


Rickettsia prowazeki prowazeki 

prowazeki typhi 

Dermacentroxenus pediculi 




Human typhus 
Murine typhus 
Tsutsugamushi disease 

Trench fever 

Rocky Mountain spotted fever 

Fi&vre boutonneuse (variety of 

spotted fever) 

Q fever 

Morphology, structure and composition of rickettsiae. The avail- 
able information concerns mainly the pathogenic rickettsiae (278). 
The type organism, Rickettsia prowazeki, the etiological agent of epi- 
demic typhus, is generally in the form of diplobacilli, 0.3 X 0.6 /* or 
somewhat longer. When in a stage of rapid growth, the bacillary 
forms may be in chains resembling those of streptococci. Their size 
makes them generally nonfiltrable, although occasionally some indi- 
viduals come through bacteria-retaining filters. They are visible in the 
microscope and resemble in stainability the Gram-negative bacteria, 
possibly with a somewhat lower affinity for basic dyes. They are 
stained with Giemsa solution or by special methods. 

Large amounts of typhus rickettsiae can be obtained either from the 
yolk sac of the chick embryo or from the intestine of infected lice. The 
yolk-sac homogenates can be purified by differential centrifugation to 
yield relatively pure suspensions of rickettsiae. These contain about 
12% nitrogen, 1% phosphorus, and about 13% lipids, a composition 

366 The Rickettsiae 

resembling that of many Gram-negative bacteria (147). They con- 
tain DNA and possibly also RNA (145; 608), the RNA being relatively 
less abundant than in most bacteria. In the electron microscope, where 
the rickettsiae show all division forms common to short bacilli (674), 
there is a clearly visible cell wall surrounded by capsular material 
(596). The capsule may be the source of the soluble antigens, which 
play an important role in the serological diagnosis of rickettsiae. The 
soluble antigens consist of particles sedimentable at medium speed in 
the ultracentrifuge and contain protein and polysaccharides, like sur- 
face antigens of bacteria. 

Other pathogenic rickettsiae fit, with minor variations, the descrip- 
tion given above for Rickettsia prowazeki. Dermacentroxenus rickettsi, 
the agent of spotted fever, often consists of pointed diplococci re- 
sembling the cells of pneumococcus. Rickettsiae very seldom show 
any forms as large as those of most bacteria, although some of the 
small Gram-negative bacilli, such as Pasteurella tularensis and P. pestis, 
come close to rickettsiae in size. 

An important fact, and one that sharply differentiates between 
rickettsiae and viruses, is that rickettsiae possess metabolic systems 
such as have never been demonstrated in any virus. Purified prepara- 
tions of typhus rickettsiae oxidize glutamate, pyruvate, and succinate 
in vitro, and produce aspartate from glutamate ( 98 ) . Apparently, the 
rickettsiae oxidize the glutamate via the tricarboxylic acid cycle, like 
mammalian tissue and several bacteria, and can presumably utilize 
the energy liberated in the oxidations. They also appear to possess a 
transaminating system. 

Serological reactions and vaccines. The serology of pathogenic 
rickettsiae has been thoroughly investigated. Antiserum, either pro- 
duced experimentally against rickettsiae or obtained from convalescent 
individuals, contains antibodies detectable by a variety of reactions 
(694). Neutralization of infectivity is tested by injection either in the 
yolk sac of the chick embryo or intraperitoneally in the guinea pig. 
The rickettsial bodies are agglutinated by antibody; the agglutination 
reaction, although quite sensitive and diagnostically valuable, since 
agglutinating antibodies appear relatively early in the course of disease, 
has the disadvantage of requiring large amounts of rickettsial antigen. 
A more convenient test is complement fixation. At least 80$ of the 
complement-fixing antigens of JR. prowazeki are in the form of soluble 
substances. These are less specific than some of the antigens present 
in the rickettsial bodies. For example, the soluble antigens from epi- 
demic and endemic (murine) typhus react equally with antisera against 

CH, 19 Serological Reactions 367 

either of the corresponding rickettsiae. Differential reactivity can be 
revealed by chemical treatments of the soluble antigens (149). Inside 
the rickettsial bodies we find both the type-specific and the group- 
specific antigens. The specific soluble antigens appear to play an im- 
portant part in causing immunity, since their presence in a vaccine is 
necessary for satisfactory protection against infection. 

Pathogenic rickettsiae also contain specific toxins. These, adminis- 
tered in sufficient dose, can kill animals in a few hours, producing 
some of the pathological changes observed in the course of the rickett- 
sial infections (297). Serum neutralization tests can distinguish be- 
tween the toxins of the epidemic and of the endemic typhus rickettsiae. 

For vaccines, large amounts of rickettsiae are prepared by a variety 
of methods (154) from the yolk sac of the chick embryo, or, for the 
agent of Rocky Mountain spotted fever, from ground bodies of ticks. 
With typhus rickettsiae, the infected yolk sacs are collected at the time 
of maximal rickettsial growth (7-8 days after inoculation with the 
inocula generally used); they are homogenized in a sterile buffer, and 
ether is added. Freezing and thawing results in the separation of an 
ether phase from the aqueous phase, which is rich in antigens. The 
water phase is collected and used as vaccine, after addition of some 
preservative substance. For use as diagnostic antigens, the rickettsiae 
can then be washed by differential centrifugation; the washing also 
removes most of the group antigens. 

Until the introduction of the egg vaccines, remarkable success had 
been obtained in typhus fever with the Weigl-type vaccines obtained 
by inoculating body lice rectally with rickettsial preparations and 
allowing the lice to feed on immune individuals. Several days later 
each louse contains as many as 100 million rickettsiae, as counted in 
microscopic preparations. The vaccine is prepared by emulsifying the 
intestine of the infected lice in phenolized saline solution. A Rocky 
Mountain spotted fever vaccine can be prepared from infected, en- 
gorged ticks homogenized in phenol or formalin solution. The crude 
homogenate is allowed to age for several days, after which it is diluted 
and clarified. 

The results of vaccination with dead rickettsiae have been remark- 
able. During World War II, the incidence of typhus among vaccinated 
troops was exceedingly low. In Naples, Italy, an epidemic of typhus 
that developed after the Allied occupation in 1943-1944 was brought 
to an abrupt end by vigorous delousing measures in the population. 
The vaccination of troops kept the epidemics from spreading to the 

368 The Rickettsiae 

armies. Such a feat was achieved for the first time in the history of 
this dreaded disease (53; 696). 

The Weil-Felix reaction. A remarkable serological reaction of 
rickettsiae, and one not yet fully understood, although widely utilized 
for diagnostic purposes, is the so-called Weil-Felix reaction (673). 
Sera of patients or convalescents of several rickettsial diseases have 
the specific property of agglutinating the cells of certain strains of the 
bacterium Proteus vulgaris. The classic strain is Proteus OX-19, which 
is agglutinated by sera from cases of endemic or epidemic typhus, 
less well by sera from Rocky Mountain spotted fever. The latter sera 
give better reaction with cells of a different strain, Proteus OX-2. 
Patients with tsutsugamushi disease have agglutinins for a different 
strain, Proteus OX-K. The agglutination involves the so-called O 
(somatic) antigens of the bacterium. No Proteus agglutinin has been 
found in the sera of patients with Q fever. 

It is likely that the Proteus organisms and the rickettsiae, although 
not closely related, possess cross-reacting antigens because of acci- 
dental structural similarity between some of the respective antigenic 
determinants. Many antigenic fractions have been isolated from the 
rickettsial bodies, and it seems that only one of the antigens is con- 
cerned with the production of the antibodies tested in the Weil-Felix 

The reproduction of rickettsiae in their vertebrate hosts. The inter- 
action between pathogenic rickettsiae and the cells of their vertebrate 
hosts has been studied only to a limited extent. Titration of rickettsiae 
can be performed by end-point methods in the guinea pig, in eggs, or 
in tissue cultures. A yolk-sac suspension may contain as much as 10 
infectious units per ml. In the chick embryo, rickettsiae grow in prac- 
tically all tissues. If inoculated in the yolk sac, they proliferate well 
in the cells that line the yolk sac and many of them are liberated into 
the yolk. There are differences in the intracellular location of different 
rickettsiae. Those of the typhus group are generally intracytoplasmic; 
those of the Rocky Mountain fever group are mainly intranuclear. 
They do not seem to form characteristic inclusions exhibiting elaborate 
cycles of the type described for several virus diseases, for example, in 
psittacosis (89). 

Several observations point to the fact that the reproduction of 
rickettsiae in the host cells takes place mainly when the cell metabo- 
lism is quite low. For example, in tissue cultures the multiplication of 
rickettsiae takes place mainly after the culture has been incubated for 
several days, when the respiratory activity of the cells has almost 

CH. 19 Rickettsiae and Arthropods 369 

ceased (697). Rickettsia prowazeki in the chick embryo grows best at 
a temperature of 32 C, where the host metabolism is low. It can be 
made to grow at 40 C in the chick embryo by the addition of cyanide, 
which depresses the metabolic rate of the cells. Rickettsial multiplica- 
tion is reduced by certain dyes which increase the rate of oxidations 
(278). Riboflavine-deficient rats are more susceptible to rickettsial 
infections than normal rats. All these observations suggest a greater 
proliferation of the parasite in conditions of metabolic deficiency of 
the host. 

In the course of studies directed toward perfecting a chemothera- 
peutic approach to rickettsial diseases, the remarkable fact was ob- 
served that, whereas sulfonamides actually increase the proliferation 
of rickettsiae and increase the severity of the diseases, para-aminoben- 
zoic acid, the normal metabolite whose utilization the sulfonamides 
inhibit, is a specific inhibitor of rickettsial reproduction (614). The 
mechanism of this rickettsiostatic action of para-aminobenzoic acid is 
unknown. This substance is a growth factor for several organisms and 
is probably an essential intermediary metabolite for all organisms, as 
a constituent of folic acid. It is possible that it plays a role in certain 
metabolic activities of the cells, which beyond a certain level become 
incompatible with rickettsial reproduction. 

Among the antibiotics, penicillin is somewhat active in suppressing 
rickettsial growth, but the most successful agents are chloromycetin 
and aureomycin, which may be said to have brought the problem of 
the therapy of rickettsial diseases under adequate control (162). 

A study of the pathology of rickettsial diseases in their vertebrate 
hosts yields little information about the biological properties of the 
agents. Most of the damage to the host depends on the invasion of 
the cells that surround small blood vessels and capillaries ( endothelial 
cells and smooth muscle fibers). This explains the occurrence of 
hemorrhagic skin rashes, including peculiar scrotal lesions observed 
in typhus infection of guinea pigs and also of men. In the central 
nervous system perivascular nodules are observed (682). 

One of the rickettsial diseases of man, Q fever, does not present these 
capillary lesions. Concomitantly, Coxiella burneti, the agent of Q 
fever, appears to be somewhat different from the other agents of the 
group, mainly because of its ability to grow extracellularly and not 
only as a strict intracellular parasite. 

Rickettsiae and arthropods. One of the most interesting problems 
posed by rickettsiae is that of their relation with their arthropod hosts. 
It is generally agreed that, with few exceptions, rickettsiae are only 

370 The Rickettsiae 

accidental parasites of vertebrates and undergo most of their evolution 
in the body of arthropods (626). 

Many insects and arachnids contain some symbionts, which resemble 
a variety of free-living organisms, yeasts, bacteria, or rickettsiae (626). 
There is no definite evidence that the symbionts that resemble bacteria 
("bacteroids") have ever been cultivated outside the arthropod body; 
therefore, their relationship with bacteria is suggested mainly on the 
basis of morphological similarity. The bacteroids have very definite 
and specific relations with the body of their host, relations that differ 
from insect to insect but that indicate very precise mechanisms for the 
maintenance of the symbiotic relation. In most insects, the symbionts 
are contained in special organs, the mycetomes. These appear to be 
formed as a response to the stimulus produced by the symbionts them- 
selves, which enter the egg at an early stage of its development. 

A number of arthropods, mainly ticks and mites, carry symbionts 
which closely resemble in size and properties the rickettsial agents of 
vertebrate diseases. In general, these rickettsiae are not carried in 
well-formed mycetomes; they are found mainly in the cells that line 
the gut. From here they may reach the genital tract, infect the egg, 
and be transmitted from generation to generation. 

The arthropod hosts do not show any abnormal symptom deriving 
from the presence of the symbionts. Actually, it is generally supposed 
that the symbionts, at least the bacteroids of the roaches, which have 
been investigated more than other symbionts, may perform some func- 
tion useful to the host. Some roaches die if the bacteroids are elimi- 
nated by the administration of agents such as penicillin (102). Some 
other insects, however, may be deprived of their symbionts without 
damage (382). 

The rickettsiae pathogenic for man present various degrees of rela- 
tion with arthropod hosts, and these relations provide the basis for 
what seems a sensible picture of the recent evolution of these micro- 
organisms (34; 113): 

1. Dermacentroxenus rickettsi, the agent of Rocky Mountain spotted 
fever, is a natural inhabitant of the gut of many ticks, where it is 
transmitted from generation to generation through the egg. It is 
occasionally inoculated by the ticks into horses and other animals, 
including man. These vertebrate hosts do not play any essential part 
in the life cycle of the organism, except that some of them, especially 
wild horses, serve to transmit the rickettsiae from infected to non- 
infected ticks. The rickettsia is only infectious if injected by a tick 
recently engorged with blood. Rickettsiae from nonengorged ticks can 

CH, 19 Rickettsiae and Arthropods 371 

even be used as a vaccine without previous inactivation, with little 
danger of infection. This unexplained property (517) may be of 
importance for the survival of the rickettsiae in nature, because it 
may help them propagate in wild animals without causing a fatal 

2. Rickettsia orientalis, the agent of scrub typhus, is found in mites 
that infest wild rats. In the mite the rickettsia is transmitted from one 
generation to the next through the egg. The rats represent an im- 
portant reservoir. Only accidentally is man infected by the bite of 
mites or of rat fleas. 

3. In the endemic (murine) typhus, the rat has become the main 
reservoir of R. prowazeki var. typhi and responds to infection with an 
almost symptomless disease. The rat fleas transmit the disease from 
rat to rat and occasionally to man, but do not have any intrinsic bio- 
logical relation with the rickettsiae, nor do they act as a reservoir. 

4. The epidemic typhus presents a different situation. The agent, 
R. prowazeki, has a cycle restricted to man and to the human louse. 
The louse receives the rickettsiae from man and becomes infected. It 
transmits the rickettsiae by excreting them in its feces. The rickettsiae 
enter the human skin through scratches provoked by the louse's bites. 
The louse generally dies as a result of the infection. Neither man nor 
louse are "natural" hosts; they are both very susceptible. In the ab- 
sence of a known reservoir of the usual kind, the disease is only main- 
tained in certain areas because of its widespread occurrence. Some 
individuals probably act as carriers. The rickettsiae survive for a 
long time in dry feces of lice. These relations explain the great im- 
portance of delousing in the prevention and control of epidemics. A 
tremendous improvement has been brought about by the introduction 
of powerful insecticides such as DDT. 

5. The agent of Q fever, C. burneti y multiplies in ticks and is ex- 
creted on cattle. In man the disease is often, but not exclusively, 
observed in slaughterhouse workers, who become infected by handling 
contaminated cattle hides. 

In view of this evidence, it seems reasonable to assume that the 
rickettsiae represent a group of bacteria which, early in evolution, 
established close biological relations with arthropods, first infecting 
them as pathogens, then becoming symbionts. The rickettsiae may 
originally have resembled the symbiotic bacterioids and may have later 
become modified within the arthropod body. Occasionally, a rickettsia 
finds its way into a vertebrate host and, if endowed with pathogenic 

372 The Rickettsiae 

ability, may produce a disease. This is generally not important in the 
life history of the rickettsiae. Exceptionally, however, a rickettsia may 
find in the vertebrate host conditions favorable for survival, as in 
murine typhus in the rat. From the vertebrate it may pass into a 
secondary arthropod, such as the rat flea. This may create a new 
vertebrate-arthropod cycle and short circuit the original arthropod 
host, which may entirely disappear from the epidemiological picture. 
By shifting from one host to another, the point may be reached where 
the cycle consists of man plus a secondary or tertiary vector. This 
may be so in epidemic typhus, which has probably derived from the 
murine type by replacement of the rat-flea cycle with the man-louse 
cycle. The disease might finally become strictly human. In fact, 
Brill's disease, a form of epidemic typhus, appears without louse in- 
festation and may be transmitted directly by human carriers. 

As already mentioned, the typhus rickettsia kills the body louse. 
Thus, the relation between a rickettsia and an arthropod is not always 
one of symbiosis or innocuous commensalism. A rickettsialike organ- 
ism, Coxiella popillia, is responsible for the blue disease of Japanese 
beetles (199). The organism fills the diseased cells of the larvae and is 
found in the larval blood. Interestingly enough, the diseased cells also 
contain intranuclear crystals. The formation of intranuclear crystals 
may be a rather general reaction to a variety of infections of insect 

It is interesting to inquire whether there are any other microorgan- 
isms that may share some of the properties of rickettsiae (518). The 
closest relative is probably the bacterium Pasteurella tularensis, the 
agent of tularemia. This tiny, Gram-negative bacterium, not appre- 
ciably larger than some of the rickettsiae, is a normal inhabitant 
of the gut of ticks. In these animals it can be transmitted through 
the egg. The only important difference between P. tularensis and 
rickettsiae is that tularensis is cultivable in bacteriological cell-free 

This criterion may not be very significant. If we conceive of 
rickettsiae as bacteria that have become intracellular parasites of 
arthropods, we may well expect to find that some rickettsiae are still 
able to grow in nonliving media. It would indeed not be surprising 
if several rickettsiae could soon be cultivated in cell-free media. All 
evidence on the metabolic relation between rickettsiae and their hosts 
suggests that the host acts more as a supplier of substrates than of 
metabolic machinery. This is at variance with the relation between 
viruses and host cells. Interesting in. this connection is the fact that 

CH. 19 Rickettsiae and Arthropods 373 

some rickettsiae have been reported to multiply, within their arthropod 
hosts, not only inside the cells of the intestinal tract but also at the 
surface of such cells, within the intestinal lumen. 

Another organism that has been considered a close relative of 
rickettsiae is Bartonella bacilliformis, the agent of the South American 
disease, Orroya fever, and of its milder form, verruga peruana. The 
natural reservoirs of this organism have not been established. Its 
relation to the arthropod host, the sand-fly, t has not yet been clarified. 

Bibliography-Author Index 

Note: Final numbers in square brackets refer to text pages. 

1. Ackermann, W. W. (1951) Concerning the relation of the Krebs cycle 

to virus propagation. 7. Biol. Chem. 189: 421-428 [229] 

2. Ackermann, W. W. ( 1951 ) The role of Z-methionine in virus propagation. 

/. Exp. Med. 93: 337-343 [146, 230] 

3. Ackermann, W. W., and Francis, T., Jr. ( 1950) Some biochemical aspects 

of herpes infection. Proc. Soc. Exp. Biol Med. 74: 123-126 [238] 

4. Ackermann, W. W., and Kurtz, H. (1952) The relation of herpes virus 

to host cell mitochondria. 7. Exp. Med. 96: 151-157 [237] 

5. Adams, M. H. (1948) Surface inactivation of bacterial viruses and of pro- 

teins. 7. Gen. Physiol. 31:417-431 [143] 

6. Adams, M. H. (1949) The stability of bacterial viruses in solutions of 

salts. J. Gen. Physiol. 32: 579-594 [138, 141, 142] 

7. Adams, M. II. ( 1949) The calcium requirement of coliphage T5. 7. Bact. 

62:505-516 [168] 

8. Adams, M. H. (1951) The hybridization of coliphage T5 and Salmonella 

phage PB. 7. Immunol 67: 313-320 [293] 

9. Adams, M. H. (1952) Classification of bacterial viruses: Characteristics 

of the T5 species and of the T2, C16 species. 7. Bact. 64: 387-396 
[17, 18, 19] 

10. Adams, M. H., and Lark, G. (1950) Mutation to heat resistance in coli- 

phage T5. 7. Immunol. 64: 335-347 [293] 

11. Alexander, H., and Leidy, G. (1951) Determination of inherited traits of 

H. influenzae by desoxyribonucleic acid fractions isolated from type- 
specific cells. 7. Exp. Med. 93: 345-359 [126, 351] 

12. Altenburg, E. (1946) The "viroid" theory in relation to plasma-genes, 

viruses, cancer and plastids. Am. Naturalist 80: 559-567 [361] 

13. American Public Health Association. (1948) Diagnostic procedures for 

virus and rickettsial diseases. New York [133] 
13a. Anderson, E. S., and Felix, A. ( 1953) Degraded Vi strains and variation 

in Vi-phage II of Salmonella typhi. 7. Gen. Microbiol 8:408-420 

13b. Anderson, E. S., and Felix, A. (1953) The Vi-type determining phages 

carried by Salmonella typhi. 7. Gen. Microbiol. 9: 65-88 [207] 

14. Anderson, K. ( 1942 ) Dual virus infections of single cells. Am. 7. Path. 

18: 577-583 [238, 284] 

15. Anderson, S. G. (1947) The mechanism of the union between influenza 

virus and susceptible cells. Austral. 7. Set. 10: 83 [258] 

16. Anderson, S. G., Burnet, F. M., Fazekas de St. Groth, S., McCrea, J. F., and 

Stone, J. D. ( 1948 ) Mucins and mucoids in relation to influenza virus 
action. Austral. J. Exp. Biol. Med. Set. 26: 403-411 [258] 


376 Bibliography 

17. Anderson, T. F. (1946) Morphological and chemical relations in viruses 

and bacteriophages. Cold Spring Harbor Symp. Quant. Biol. 11: 1-13 

18. Anderson, T. F. (1948) The growth of T2 virus on ultraviolet-killed host 

cells. J. Bact. 56: 403-410 [179] 

19. Anderson, T. F. (1949) The reaction of bacterial viruses with their host 

cells. Bot. Rev. 15: 464-505 [159] 

20. Anderson, T. F. (1950) Destruction of bacterial viruses by osmotic shock. 

/. Appl. Phys. 21: 70 [65] 

21. Anderson, T. F., and Doermann, A. H. (1952) Sonic reactivation of anti- 

serum-neutralized bacteriophage T3. J. Bact. 63: 291-292 [123] 

22. Anderson, T. F., and Stanley, W. M. (1941) A study by means of the 

electron microscope of the reaction between tobacco mosaic virus and 
its antiserum. /. Biol. Chem. 139: 339-344 [67] 

23. Andervont, H. B. (1945) The milk influence in the genesis of mammary 

tumors. In: A Symposium on Mammary Tumors in Mice. A.A.A.S., 
Washington [328, 340] 

24. Andrewes, C. H. (1929) Virus III in tissue cultures. II, III. Br. J. Exp. 

Path. 10: 273-280 [227, 229, 281] 

25. Andrewes, C. H. ( 1934 ) Viruses in relation to the aetiology of tumours. 

Lancet (2) 227:63-69; 117-123 [334, 335] 

26. Andrewes, C. H. (1936) Evidence for the presence of virus in a nonfilter- 

able tar sarcoma of the fowl. /. Path. Bact. 43: 23-33 [338] 

27. Andrewes, C. H. (1942) Interference by one virus with the growth of an- 

other in tissue culture. Br. J. Exp. Path. 23: 214-220 [267, 283] 

28. Andrewes, C. H. (1952) The place of viruses in nature. Proc. Roy. Soc. 

(London) B, 139:313-326 [272] 

29. Andrewes, C. H., and Elford, W. J. (1933) Observations on antiphagc 

sera. I. The "percentage law." Br. J. Exp. Path. 14:367-376 [122] 

30. Andrewes, C. II., Ahlstrom, C. G., Foulds, L., and Gye, W. E. (1937) 

Reaction of tarred rabbits to the infectious fibroma virus ( Shope ) . Lancet 
(2) 233:893-895 [338] 

31. Armstrong, C., Wallace, J. J., and Ross, L. (1940) Lymphocytic chorio- 

meningitis. Gray mice, Mus musculus, a reservoir for the infection. U. S. 
Pub. Health Rep. 55: 1222-1229 [38, 309, 340] 

32. Avery, O. T., MacLeod, C. M., and McCarty, M. (1944). Studies on the 

chemical nature of the substance inducing transformation of pneumococ- 
cal types. Induction of transformation by a desoxyribonucleic acid frac- 
tion isolated from pneumomoccus type III. J. Exp. Med. 79: 137-158 
[100, 351] 

33. Backus, R. C., and Williams, R. C. (1950) The use of spraying methods 

and volatile suspending media in the preparation of specimens for electron 
microscopy. /. Appl. Phys. 21: 11-15 [211] 

34. Baker, A. C. (1943) The typical epidemic series. Am. }. Trop Med. 

23: 559-567 [364, 370] 

35. Bald, J. G. (1937) The use of numbers of infections for comparing the 

concentration of plant 'virus suspensions. IV. Austral. J. Exp. Biol. Med. 
Set. 15: 211-220 [54] 

Bibliography 377 

36. Banfield, W. G., Bunting, H., Strauss, M. J., and Melnick, J. L. (1951) 

Electron micrographs of thin sections of molluscum contagiosum. Proc. 
Soc. Exp. Biol Med. 77: 843-847 [234] 

37. Bang, F. B. (1948) Studies on Newcastle disease virus. III. /. Exp. 

Med. 88:251-266 [82] 

37a. Bang, F. B., and Gye, G. O. (1952) Comparative susceptibility of cul- 
tured cell strains to the virus of Eastern equine encephalomyelitis. Bull 
Johns Hopkins Hosp. 91: 427-461 [227] 

38. Barnard, J. E. (1932) Discussion on the microscopy of the filterable 

viruses. /. Roy. Microsc. Soc. 52: 233-235 [60] 

39. Barnard, J. E. (1935) Microscopical evidence of the existence of sapro- 

phytic viruses. Br. J. Exp. Path. 16: 129-133 [4] 

40. Barton-Wright, E., and McBain, A. (1932) Studies in the physiology of 

the virus diseases of the potato: a comparison of the carbohydrate metabo- 
lism of normal with that of leaf-roll potatoes. Trans. Roy. Soc. Edinburgh 
57: 309-349 [218] 

41. Bauer, D. J. (1953) Metabolic aspects of virus multiplication. In: The 

Nature of Virus Multiplication, Cambridge University Press [84, 108, 

42. Baumgartner, G. (1935) Infektionversuche mit isolierten oxychromatischen 

Einschlussen des Herpes. Schweiz. med. Wochenschr. 16: 759-760 

43. Bawden, F. C. ( 1950 ) Plant Viruses and Virus Diseases, 3rd Ed. Waltham, 

Mass., Chronica Botanica [3, 6, 16, 17, 93, 94, 121, 137, 141, 215, 
216, 218, 315] 

44. Bawden, F. C., and Kassanis, B. (1945) The suppression of one plant 

virus by another. Ann. Appl. Biol. 32: 52-57 [278] 

45. Bawden, F. C., and Kassanis, B. (1949) Some effects of host nutrition on 

the multiplication of viruses. Ann. Appl. Biol. 37: 215-228 [219] 

46. Bawden, F. C., and Kassanis, B. (1951) Serologically related strains of 

potato virus Y that are not mutually antagonistic in plants. Ann. Appl. 
Biol 38: 402-410 [278] 

47. Bawden, F. C., and Kleczkowski, A. (1945) Protein precipitation and 

virus inactivation by extracts of strawberry plants. J. Pomol. 21: 2-7 

48. Bawden, F. C., and Nixon, H. L. ( 1953) Quoted by Bawden and Pirie in: 

The Nature of Virus Multiplication, Cambridge University Press [213] 

49. Bawden, F. C., and Pirie, N. W. ( 1950) Some factors affecting the activa- 

tion of virus preparations made from tobacco leaves infected with a tobacco 
necrosis virus. /. Gen. Microbiol. 4: 464-481 [110, 128, 215] 

50. Bawden, F. C., and Pirie, N. W. (1950) The varieties of macromolecules 

in extracts from virus-infected plants. In: Delbruck, Viruses 1950, Pasa- 
dena, Cal. Inst. Technol. Bookstore [86] 

51. Bawden, F. C., and Roberts, F. M. (1948) Photosynthesis and predis- 

position of plants to infection with certain viruses. Ann. Appl Biol 
35: 418-428 [219] 

52. Bawden, F. C., Kassanis, B., and Nixon, H, L. (1950) The mechanical 

transmission and some properties of potato paracrinkle virus. J. Gen. 
Microbiol 4: 210-219 [34, 349] 

378 Bibliography 

53. Bayne-Jones, S. (1948) Epidemic typhus in the Mediterranean area dur- 

ing World War II. In: The Rickettsial Diseases of Man. A.A.A.S., Wash- 
ington [368] 

54. Beale, H. P. (1937) Relation of Stanley's crystalline tobacco- virus protein 

to intracellular crystalline deposits. Contrib. Boyce Thompson Inst. 
8: 413-431 [215] 

54a. Beale, H. P. (1947) The interference phenomenon between the rib- 
grass and tobacco-mosaic viruses in bean. Phytopath. 37: 847 [277] 

55. Beale, H. P., and Jones, C. R. ( 1951 ) Virus diseases of tobacco mosaic 

and potato yellow dwarf not controlled by certain antibiotics. Contrib. 
Boyce Thompson Inst. 16: 395-407 [219] 

56. Beard, J. W. (1948) Purified animal viruses. J. Immunol 58:49-108 

[105, 106] 

57. Beard, J. W., Bryan, W. R., and Wyckoff, R. W. G. ( 1939) The isolation 

of the rabbit papilloma virus protein. /. Infectious Diseases 65: 43-52 
[327, 336] 

58. Beijetfnck, M. W. (1899) Uber ein Contagium vivum fluidum als Ursachc 

der Fleckenkrankheit der Tabaksblatter. Zentr. Bakt., Parasit., I/, 5: 27-33 

59. Bennett, C. W. (1935) Studies on properties of the curly top virus. 

J. Agr. Res. 50: 211-241 [33] 

60. Bennett, C. W. (1937) Correlation between movement of the curly top 

virus and translocation of food in tobacco and sugar beet. J. Agr. Res. 
54: 479-502 [217] 

61. Bennett, C. W. (1940) Acquisition and transmission of viruses by dodder 

(Cuscuta subinclusa). Phytopath. 30: 2 [34] 

61a. Bennett, C. W. (1940) The relation of viruses to plant tissues. Bot. 
Rev. 6: 427-473 [216] 

62. Bennett, C. W. (1944) Latent virus of dodder and its effect on sugar 

beet and other plants. Phytopath. 34: 77-91 [37] 

63. Bennett, C. W., and Wallace, H. E. (1938) Relation of the curly top 

virus to the vector Eutettix tenellus. J. Agr. Res. 56: 31-51 [317] 

64. Benzer, S. (1952) Resistance to ultraviolet light as an index to the repro- 

duction of bacteriophage. /. Bact. 63: 59-72 [156, 195, 196] 

65. Benzer, S., et al. (1950) A syllabus of procedures, facts, and interpreta- 

tions in phage. In: Delbriick, Viruses 1950, Pasadena, Gal. Inst. Technol. 
Bookstore [7] 

66. Bergold, G. (1947) Die Isolierung des Polyeder- Virus und die Natur der 

Polyeder. Z. Naturforsch. 2b: 122-143 [83, 110] 

67. Bergold, G. (1948) Uber die Kapselvirus-Krankheit. Z. Naturforsch. 

3b: 338-342 [83] 

68. Bergold, G. H. (1950) The multiplication of insect viruses as organisms. 

Canad. J. Res., E, 28: 5-11 [83] 

68a. Bergold, G. H. (1953) Insect viruses. Adv. Virus Research 1:91-139 
[19, 83, 110, 237] 

69. Bernal, J. D., and Fankuchen, I. ( 1941 ) X-ray and crystallographic studies 

of plant virus preparations. J. Gen. Physiol 25: 111-165 [93, 94, 95] 

70. Bernal, J. D., Fankuchen, I., and Riley, D. P. (1938) Structure of the 

crystals of tomato bushy stunt virus preparations. Nature 142: 1075 

Bibliography 379 

71. Bernkopf, H. (1950) Study of infectivity and hemagglutination of in- 

fluenza virus in deembryonated eggs. /. Immunol. 65: 571-583 [265] 

72. Bernkopf, H., and Kligler, I. J. (1940) Characteristics of a fixed rabies 

virus cultivated on developing chick embryos. Proc. Soc. Exp. Biol. Med. 
45: 332-335 [245, 246] 

73. Berry, G. P., and Dedrick, H. M. (1936) A method for changing the 

virus of rabbit fibroma (Shope) into that of infectious myxomatosis 
(Sanarelli). J. Bad. 31: 50 [302] 

74. Bertani, G. (1951) -Studies on lysogenesis. I. /. Bact. 62:293-300 


75. Bertani, G. (1953) Infection bacteriophagique secondaire des bacteries 

lysogcnes. Ann. Inst. Pasteur 84: 273-280 [203, 277] 

76. Bertani, G., and Weiglc, J. J. ( 1952) Host-controlled variation in bacterial 

viruses. /. Bact. 65: 113-121 [197, 198, 295, 333] 

77. Bevoridge, W. I. B., and Burnet, F. M. (1946) The cultivation of viruses 

and rickettsiae in the chick embryo. Med. Res. Council Spec. Rep. No. 
256 [42, 239, 241, 247, 249] 

78. Billingham, R. E., and Medawar, P. B. (1948) Pigment spread and cell 

heredity in guinea-pig's skin. Heredity 2: 29-47 [359] 

78a. Bird, F. T. ( 1949) Tumours associated with a virus infection in an insect. 
Nature 163: 777-778 [325] 

79. Bittner, J. J. (1936) Some possible effects of nursing on the mammary 

gland tumor incidence in mice. Science 84: 162 [328] 

80. Bittner, J. J. (1941) Changes in the incidence of mammary carcinoma 

in mice of the A stock. Cancer Res. 1: 113-114 [330] 

81. Black, L. M. (1941) Specific transmission of varieties of potato yellow 

dwarf virus by related insects. Am. Potato J. 18: 231-233 [315] 

82. Black, L. M. ( 1941 ) Further evidence for multiplication of the aster-yellow 

virus in the aster leafhopper. Phytopath. 31: 120-125 [317] 

83. Black, L. M. (1945) A virus tumor disease of plants. Am. J. Bot. 

32: 408-415 [7, 29, 216, 317, 330] 

84. Black, L. M. (1950) A plant virus that multiplies in its insect vector. 

Nature 166: 852-853 [317] 

85. Black, L. M., and Brakke, M. K. (1952) Multiplication of wound-tumor 

virus in an insect vector. Phytopath. 42: 269-273 [318] 

86. Black, L. M., Morgan, C., and Wyckoff, R. W. G. ( 1950) Visualization of 

tobacco mosaic virus within infected cells. Proc. Soc. Exp. Biol. Med. 
73: 119-122 [213] 

87. Blanc, G. (1948) Dengue. In: Levaditi and Lepine, Les ultravirus des 

maladies humaines, 2nd Ed., Paris, Maloine [311] 

88. Blanc, G., and Caminopetros, J. (1929) Duree de conservation du virus 

de la dengue chez les Stegomyas. L'influence de la saison froide sur le 
pouvoir infcctant. C. r. acad. sci. 188: 1273-1275 [312] 

89. Bland, J. O. W., and Canti, R. G. ( 1935) The growth and development of 

psittacosis virus in tissue cultures. J. Path. Bact. 40: 231-241 [58, 
232, 233, 247, 368] 

90. Bland, J. O. W., and Robinow, C. F. (1939) The inclusion bodies of 

vaccinia and their relationship to the elementary bodies studied in cul- 
tures of the rabbit's cornea. J. Path. Bact. 48: 381-403 [233] 

380 Bibliography 

91. Bodian, D. (1952) A reconsideration of the pathogcnesis of poliomyelitis. 

Am. J. Hyg. 55: 414-438 [308] 

92. Bodian, D., and Howe, H. A. (1941) The rate of progression of polio- 

myelitis virus in nerves. Bull Johns Hopkins Hosp. 69: 79-85 [308] 

93. Bodian, D., and Howe, H. A. ( 1947) The significance of lesions in periph- 

eral ganglia in chimpanzee and in human poliomyelitis. J. Exp. Med. 
85: 231-242 [308] 

94. Born, H. J., Lang, A., and Schramm, G. ( 1943) Markierung von Tabakmo- 

saikvirus mit Radiophosphor. Arch. ges. Virusforsch. 2: 461-479 [210] 

95. Borrel, A. (1903) Epithelioses infectieuses et epitheliomas. Ann. Imt. 

Pasteur 17: 81-84 [324] 

96. Borrel, A. ( 1904) Sur les inclusions de repithelioma contagieux des oiseaux. 

C. r. soc. biol 57: 642-643 [232] 

97. Borrel, A. ( 1907) Le probleme du cancer. Bull. Inst. Pasteur 5: 497-512, 

545-562, 593-608, 641-662 [324] 

98. Bovarnich, M. R., and Miller, J. C. (1950) Oxidation and transamination 

of glutamate by typhus rickettsiae. 7. Biol. Chem. 184: 661-676 [366] 
98a. Bowen, G. (1953) Kinetic studies on the mechanism of photoreactiva- 
tion of bacteriophage T2 inactivated by ultraviolet light. Ann. Inst. 
Pasteur 84: 218-221 [195] 

99. Boyd, J. S. K. (1950) The symbiotic bacteriophages of Salmonella typhi- 

murium. J. Path. Bact. 62: 501-517 [206, 208] 

100. Boyd, J. S. K. ( 1951 ) Mutation in a bacterial virus. Nature 168: 994-995 

[203, 293] 

lOOa. Brakke, M. K. (1951) Density gradient centrif ugation : A new separa- 
tion technique. 7. Am. Chem. Soc. 73: 1847 [89] 

lOOb. Brakke, M. K., Black, L. M., and Wyckoff, R. W. G. (1951) The sedi- 
mentation rate of potato yellow-dwarf virus. Am. J. Bot. 38: 332-342 

lOOc. Brakke, M. K., Vatter, A. E., and Black, L. M. Personal communication. 

101. Braun, A. C. (1952) The crown-gall disease. Ann. N. Y. Acad. Sci. 

54: 1153-1160 [167, 339] 

lOla. Briody, B. A., and Stannard, C. ( 1951 ) Studies on vaccinia virus. I. 
The development of hemagglutinating and infective particles in the chorio- 
allantois of the chick embryo. 7. Immunol. 67: 403-411 [269] 

102. Brues, C. T., and Dunn, R. C. (1945) The effect of penicillin and certain 

sulfa drugs on the intracellular bactcroids of the cockroach. Science 
101: 336-337 [370] 

103. Bryan, W. R., and Beard, J. W. (1939) Estimation of purified papilloma 

virus protein by infectivity measurements. 7. Infectious Diseases 65: 306- 
321 [45] 

104. Bryan, W. R., and Beard, J. W. ( 1940) Host influence in the characteriza- 

tion of response to the papilloma protein and to vaccinia virus. 7. Infec- 
tious Diseases 67: 5-24 [53, 54] 

105. Bryan, W. R., and Beard, J. W. (1941) Studies on the purification and 

properties of the rabbit-papilloma-virus protein. 7. Nat. Cancer Inst. 
1:607-673 [120] 

Bibliography 381 

106. Buist, J. B. (1887) Vaccinia and Variola: A Study of Their Life History, 

London, Churchill [6, 58] 

107. Burmester, B. R. ( 1952) Studies on fowl lymphomatosis. Ann. N. Y. Acad. 

Sci. 54: 992-1003 [239, 325] 

108. Burnet, F. M. (1929) A method for the study of bacteriophage multipli- 

cation in broth. Br. J. Exp. Path. 10: 109-115 [162] 

109. Burnet, F. M. (1930) Bacteriophage activity and antigenic structure 

of bacteria. /. Path. Bact. 33: 647-664 [168] 

110. Burnet, F. M. (1933) The serological classification of Coli-dysentery 

phages. J. Path. Bact. 36: 307-318 [122] 

111. Burnet, F. M. (1933) A specific soluble substance from bacteriophages. 

Br. J. Exp. Path. 14: 100-108 [111] 

112. Burnet, F. M. (1940) The production of antibodies. Monographs from 

Walter and Eliza Hall Institute, I [132] 

113. Burnet, F. M. (1942) The rickettsial diseases in Australia. Med. J. Aus- 

tralia 2: 129-134 [370] 

114. Burnet, F. M. (1945) Virus as Organism, Cambridge, Harvard University 

Press [4, 5, 37, 271, 290, 302, 310, 319, 347] 

115. Burnet, F. M. (1948) The initiation of cellular infection by influenza 

and related viruses. Lancet (1) 254:7-11 [253, 254, 256] 

116. Burnet, F. M. (1951) Some biological implications of studies on influenza 

\iruses. Bull Johns Hopkins Hosp. 88: 119-180 [253] 

117. Burnet, F. M., and Boake, W. C. (1946) The relationship between the 

\irus of infectious ectromelia of mice and vaccinia virus. /. ImmunoL 
53: 1-13 [301] 

118. Burnet, F. M., and Bull, D. R. (1943) Changes in influenza virus asso- 

ciated with adaptation to passage in chick embryos. Austral. J. Exp. 
Biol Med. Sci. 21: 55-69 [253, 300] 

118a. Burnet, F. M., and Edney, M. ( 1951) Recombinant viruses obtained from 
double infections with the influenza A viruses MEL and Neuro-WS. 
Austral. J. Exp. Biol. Med. Sci. 39: 353-362 [267, 268] 

119. Burnet, F. M., and Lind, P. E. (1950) Haemolysis by Newcastle disease 

virus. II. Austral. J. Exp. Biol. Med. Sci. 28: 129-150 [254] 

120. Burnet, F. M., and Lind, P. E. ( 1951) A genetic approach to variation in 

influenza viruses. I-IV. J. Gen. Microbiol. 5: 46-82 [267, 302, 350] 
120a. Burnet, F. M., and McKie, M. (1929) Observations on a permanently 

lysogenic strain of B. enteritidis Gaertner. Austral. J. Exp. Biol. Med. Sci. 

6: 277-284 [30] 
120b. Burnet, F. M., Fraser, K. B., and Lind, P. E. (1953) Genetic interaction 

between influenza viruses. Nature 171: 163-165 [268] 

121. Burnet, F. M., Keogh, E. V., and Lush, D. (1937) The immunological 

reactions of the filterable viruses. Austral. J. Exp. Biol. Med. Sci. 
15:231-368 [7, 119, 122, 128, 131] 

121a. Burnet, F. M., and others. (1953) Virus and rickettsial classification and 
nomenclature. Ann. N. Y. Acad. Sci. 56: 381-622 [16] 

122. Cairns, H. J. F., Fazekas de St. Groth, S., and Edney, M. (1952) Quanti- 

tative aspects of influenza virus multiplication. I-IV. J. Immunol. 69: 
155-181 [261, 263, 286] 

382 Bibliography 

123. Carrel, A. (1925) Comparaison des macrophages normaux et des macro- 

phages transformed en cellules malignes. C. r. soc. biol. 92: 584-587 

124. Carrel, A. (1926) Some conditions of the reproduction in vitro of the 

Rous virus. /. Exp. Med. 43: 647-668 [226, 227] 

124a. Carter, W. ( 1928 ) Transmission of the virus of curly-top of sugar beets 
through different solutions. Phytopath. 18: 675-679 [317] 

125. Casals, J., Olitsky, P. K., and Anslow, R. O. ( 1951 ) A specific comple- 

ment-fixation test for infection with poliomyelitis virus. J. Exp. Med. 
94: 123-137 [134] 

126. Caspari, E. (1948) Cytoplasmic inheritance. Adv. Genetics 2:1-66 


127. Caspersson, T. (1939) t)ber die Rolle der Desoxyribosenucleinsaure bei 

der Zellteilung. Chromosoma 1: 147-156 [238] 

128. Chargaff, E. (1950) Chemical specificity of nucleic acids and mechanism 

of their enzymatic degradation. Experientia 6: 201-240 [97] 

129. Cheever, F. S., and Janeway, C. A. (1941) Immunity induced against 

the Brown-Pearce carcinoma. Cancer Res. 1: 23-27 [337] 

130. Chester, K. S. (1936) Serological tests with Stanley's crystalline tobacco- 

mosaic protein. Phytopath. 26: 715-734 [88] 

131. Chu, C. M., Dawson, I. M., and Elford, W. J. (1949) Filamentous forms 

associated with newly isolated influenza virus. Lancet ( 1 ) 256: 602-603 

132. Chu, L. W., and Morgan, H. P. (1950) Studies of the hemolysis of red 

blood cells by mumps virus. J. Exp. Med. 91: 393-402 [254] 

133. Clark, F., and Nagler, F. P. O. (1943) Haem-agglutination by viruses. 

The range of susceptible cells with special reference to agglutination by 
vaccinia virus. Austral J. Exp. Biol. Med. Set. 21: 103-106 [251] 

134. Clark, P. F. (1949) The influence of nutrition on experimental virus in- 

fection. Bact. Rev. 13: 122-127 [31, 223] 

135. Clarke, N. A. (1952) Studies on the host-virus relationship in a lyso- 

genic strain of Bacillus megaterium. II. J. Bact. 63: 187-192 [205, 

136. Claude, A. ( 1937 ) Preparation of an active agent from inactive tumor 

extracts. Science 85: 294-295 [335] 

137. Claude, A. ( 1938 ) Concentration and purification of chicken tumor I 

agent. Science 87: 467-468 [107] 

138. Claude, A. (1938) A fraction from normal chick embryo similar to the 

tumor-producing fraction of chicken tumor I. Proc. Soc. Exp. Biol. Med. 
39: 398-403 [342] 

139. Claude, A. (1939) Chemical composition of the tumor-producing fraction 

of chicken tumor I. Science 90: 213-214 [324] 

140. Claude, A., and Murphy, J. B. (1933) Transmissible tumors of the fowl. 

Physiol Rev. 13: 246-275 [325, 343] 

141. Cloudman, A. M. (1941) The effect of an extra-chromosomal influence 

upon transplanted spontaneous tumors in mice. Science 93: 380-381 
[322, 330, 348] 

142. Cohen, S. S. (1944) Cytoplasmic particles of chorio-allantoic membrane 

and their relations to purified preparations of influenza virus. Proc. Soc. 
Exp. Biol. Med. 57: 358-360 [107] 

Bibliography 383 

143. Cohen, S. S, (1947) The synthesis of bacterial viruses in infected cells. 

Cold Spring Harbor Symp. Quant. Biol 12: 35-49 [175, 178] 

144. Cohen, S. S. (1949) Growth requirements of bacterial viruses. Bact. Rev. 

13: 1-24 [176, 177, 178, 179] 

145. Cohen, S. S. (1950) Studies on commercial typhus vaccines. IV. J. 

Immunol 65:475-483 [366] 

146. Cohen, S. S., and Arbogast, R. (1950) Chemical studies in host-virus in- 

teraction. VI. /. Exp. Med. 91: 607-618 [88, 103] 

147. Cohen, S. S., and Chargaff, E. (1944) Studies on the composition of 

Rickettsia prowazeki. J. Biol. Chem. 154: 691-704 [366] 

148. Cohen, S. S., and Stanley, W. M. (1942) The molecular size and shape 

of the nucleic acid of tobacco mosaic virus. /. Biol. Chem. 144: 589-598 

149. Cohen, S. S., Chambers, L. A., and Clawson, J. R. ( 1950) Studies on com- 

mercial typhus vaccines. III. J. Immunol. 65: 465-473 [367] 

150. Cold Spring Harbor Symp. Quant. Biol 12 ( 1947) [99] 

150a. Commoner, B., and Dietz, P. M. (1952) Changes in non-protein nitrogen 
metabolism during tobacco mosaic virus biosynthesis. /. Gen. Physiol. 
35: 847-856 [210, 218] 

151. Commoner, B., and Mercer, F. (1951) Inhibition of the biosynthesis of 

tobacco mosaic virus by thiouracil. Nature 168: 113-114 [146, 210] 

152. Commoner, B., Mercer, F. L., Merrill, P., and Zimmer, A. J. (1950) 

Microanalytic determination of the rate of tobacco mosaic virus synthesis 
in tobacco leaf tissue. Arch. Biochem. 27: 271-286 [210, 211, 212] 

153. Cosslett, V. E. ( 1951 ) Practical Electron Microscopy, New York, Academic 

Press [60] 

154. Cox, H. R. (1948) The preparation and standardization of rickettsial vac- 

cines. In: The Rickettsial Diseases of Man. A.A.A.S., Washington 

155. Craigie, J. (1932) The nature of the vaccinia flocculation reaction, and 

observations on the elementary bodies of vaccinia. Br. J. Exp. Path. 
13: 259-268 [89] 

156. Craigie, J. (1946) The significance and applications of bacteriophage in 

bacteriological and virus research. Bact. Rev. 10: 73-88 [167] 

157. Craigie, J., and Felix, A. (1947) Typing of typhoid bacilli with Vi-bac- 

teriophages. Lancet (1) 252:823-826 [167, 206] 

157a. Craigie, J., and Yen, C. H. (1938) The demonstration of types of B. 
typhosus by means of preparations of Type II Vi phage. I, II. Canad. 
Publ Health J. 29: 448-463 and 484-496 [167, 206] 

158. Crawford, G. N. C., and Sanders, F. K. ( 1952) The multiplication of vac- 

cinia virus in tissue cultures of adult rabbit skin. Quart. J. Microscop. 
Sci. 93: 119-132 [269, 270] 

159. Grouse, H. V., Coriell, L. L., Blank, H., and McNair Scott, T. F. (1950) 

Cytochemical studies on the intranuclear inclusion of herpes simplex. /. 
Immunol. 65: 119-128 [237] 

160. Curnen, E. C., and Horsfall, F. L. (1946) Studies on pneumonia virus of 

mice (PVM), III. /. Exp. Med. 83: 105-132 [87] 

161. Curtis, M. R., Dunning, W. F., and Bullock, F. D. (1933) Genetic factors 

in relation to the etiology of malignant tumors. Am. J. Cancer 
17: 894-923 [323] 

384 Bibliography 

162. Cutting, W. C. ( 1949) Actions of antibiotics in vivo. Ann. Rev. Microbiol. 

3: 137-158 [223, 369] 

162a. Davenport, F. M. (1952) Toxicity of NDV for mouse lung. J. Immunol 
69: 461-470 [223, 224] 

163. Davenport, F. M., and Francis, T., Jr. ( 1951 ) A comparison of the 

growth curves of adapted and unadapted lines of influenza virus. J. Exp. 
Med. 93: 129-137 [221] 

164. DeBurgh, D., Yu, P. C., Howe, C., and Bovarnick, M. ( 1948) Preparation 

from human red cells of a substance inhibiting virus hemagglutination. 
7. Exp. Med. 87: 1-9 [258] 

164a. DeCarlo, M. R., Sarles, W. B., and Knight, S. G. ( 1953) Lysogenicity of 
Bacillus megatherium. J. Bact. 65: 53-55 [206] 

165. Delbriick, M. (1940) Adsorption of bacteriophagcs under various physio- 

logical conditions of the host. 7. Gen. Physiol 23: 631-642 [166] 

166. Delbriick, M. (1940) The growth of bacteriophage and lysis of the host. 

7. Gen. Physiol. 23: 643-660 [168] 

167. Delbriick, M. (1945) Effect of specific antisera on the growth of bacterial 

viruses. 7. Bact. 50: 137-150 [125, 164] 

168. Delbriick, M. (1945) Interference between bacterial viruses. III. 7. 

Bact. 50: 151-170 [187, 188, 273] 

169. Delbriick, M. (1946) Bacterial viruses or bacteriophagcs. Biol. Rev. 

21: 30-40 [122, 159] 

170. Delbriick, M. (1948) Biochemical mutants of bacterial viruses. 7. Bact. 

56: 1-16 [146, 166, 167, 293] 

171. Delbriick, M. (1949) Discussion of paper by Beale and Sonncborn. In: 

Unites biolQgiques donees de continuite genetique. Centre Nat. Rech. 
Scientifique, Paris [355] 

172. Delbriick, M. Viruses 1950, Pasadena, Cal. lust. Technol. Bookstore [1] 

173. Delbriick, M., and Bailey, W. T., Jr. (1946) Induced mutations in bac- 

terial viruses. Cold Spring Harbor Symp. Quant. Biol 11: 33-37 [189, 

174. Delbriick, M., and Luria, S. E. (1942) Interference between bacterial 

viruses. I. Arch. Biochem. 1:111-141 [159, 273] 

175. DeMars, R. L, Luria, S. E., Fisher, II., and Lcvinthal, C. (1953) The 

production of incomplete bacteriophage particles by the action of proflavine, 
and the properties of the incomplete particles. Ann. lm>t. Pasteur 84: 113- 
128 [84, 111, 120, 126, 127, 171, 172, 173, 199] 

176. Demerec, M. (1949) Chemical mutagens. Proc. VI II Intern. Congr. 

Genetics, Stockholm, 201-209 [338] 

177. Demerec, M., and Fano, U. (1945) Bactcriophage-resistant mutants in 

Escherichia coli. Genetics 30: 119-136 [22, 186] 

178. DeRobertis, E. E. D. P., Nowinsky, W. W., and Saez, F. A. ( 1948) Gen- 

eral Cytology, Philadelphia, Saunders [339, 354] 

178a. DeRopp, R. S. ( 1951 ) The crown-gall problem. Bot . Rev. 17: 629-670 

179. Dienes, L., and Weinberger, H. J. (1951) The L forms of bacteria. 

Bact. Rev. 15: 245-288 [4, 353] 

180. Dodd, K., Johnston, L. M., and Buddingh, G. J. ( 1938) Herpetic stomatitis. 

7. Pediatr. 12: 95-102 [38] 

Bibliography 385 

181. Doermann, A. H. (1948) Lysis and lysis inhibition with Escheriehia coli 

bacteriophage. /. Bact. 55: 257-276 [184, 292] 

182. Doermann, A. H. (1952) The intracellular growth of bacteriophages. I. 

/. Gen. Physiol. 35:645-656 [170, 171] 

183. Doermann, A. H., and Dissosway, C. F.-R. (1949) Intracellular growth 

and genetics of bacteriophage. Carnegie Inst. Wash. Yearbook 48: 170- 
176 [190] 

184. Doermann, A. H., and Hill, M. B. (1953) Genetic structure of bacterio- 

phage T4 as described by recombination studies of factors influencing 
plaque morphology. Genetics 38: 79-90 [190] 

185. Doerr, R. (1939) Die qualitative Virusnachweis. Anreichungsverfahren. 

In: Doerr and Hallauer, 2 (Ref. 187) [35] 

186. Doerr, R. (1939) Die quantitative Bestimmung der Infektiositat. In: 

Doerr and Hallauer, 2 (Ref. 187) [42] 

187. Doerr, R., and Hallauer, C. (1938-1939) Handbuch der Virusforschung, 

Wien, Springer [1] 

188. den Dooren de Jong, L. E. ( 1931) Ober Bac. megatherium und den darin 

anwesenden Bakteriophagen. Zentr. Bakt., Parasit., /, Grig. 120: 1-15 

189. Doty, P., and Steiner, R. F. ( 1950) Light scattering and spectrophotometry 

of colloidal solutions. /. Chem. Physics 18: 1211-1220 [80] 

190. Douglas, S. R., Smith, W., and Price, L. R. W. (1929) Generalized 

vaccinia in rabbits with special reference to lesions in the internal or- 
gans. /. Path. Bact. 39: 99-120 [300] 

191. DuBuy, H. G., and Woods, M. W. (1943) Evidence for the evolution 

of phytopathogcnic viruses from mitochondria and their derivatives. II. 
Phytopath. 33: 766-777 [213, 214] 

192. Duggar, B. M., and Armstrong, J. K. (1923) Indications respecting the 

nature of the infective particles in the mosaic disease of tobacco. Ann. 
Missouri Botan. Garden 10: 191-212 [350] 

193. Dulbecco, R. (1950) Experiments on photoreactivation of bacteriophages 

inactivated with ultraviolet radiation. /. Bact. 59: 329-347 [155, 194] 

194. Dulbecco, R. (1952) A critical test of the recombination theory of multi- 

plicity reactivation. /. Bact. 63: 199-207 [195] 

195. Dulbecco, R. (1952) Mutual exclusion between related phages. J. Bact. 

63:209-217 [189, 274, 275] 

196. Dulbecco, R. (1952) Production of plaques in monolayer tissue cultures 

by single particles of an animal virus. Proc. Nat. Acad. Set. 38: 747-752 
[42, 48, 226] 

197. Duran-Reynals, F. (1940) A hemorrhagic disease occurring in chicks 

inoculated with the Rous and Fujinami viruses, Yale J. BioL Med. 
13: 77-98 [331] 

197a. Duran-Reynals, F. ( 1942) The reciprocal infection of ducks and chickens 
with tumor-inducing viruses. Cancer Research 2: 343-369 [332] 

198. Duran-Reynals, F., and Shrigley, E. W. (1945) Virus infection as an 

etiologic agent of cancer. Res. Conf. on Cancer, A.A.A.S., Washington 
[31, 331] 

199. Dutky, S. R., and Gooden, E. L. (1952) Coxiella popittia, n. sp., a 

rickettsia causing blue disease of Japanese beetle larvae. /. Bact. 63: 743- 
750 [372] 

386 Bibliography 

199a. Eaton, M. D. (1952) Observations on growth of virus and the energy- 
yielding activities of the host cell. Arch. ges. Virusforsch. 5: 53-72 

200. Eaton, M. D., Magasanik, B., and Perry, M. E. (1951) Inhibition of in- 

fluenza and mumps viruses in tissue culture by basic amino acids. Proc. 
Soc. Exp. Biol Med. 77: 505-508 [230] 

201. Eddy, B. E., and Wyckoff, R. W. G. (1950) Influenza virus in sectioned 

tissues. Proc. Soc. Exp. Biol. Med. 75: 290-293 [267] 

202. Ehrlich, H. L., and Watson, D. W. (1949) Lysogenesis of Bacillus mega- 

therium. J. Bact. 58: 627-632 [205] 

203. Eklund, C. M. (1946) Human encephalitis of the Western equine type 

in Minnesota in 1941: Clinical and epidemiological study of serologically 
positive cases. Am. J. Hyg. 43: 171-193 [314] 

204. Elford, W. J. (1938) The sizes of virus and bacteriophages and methods 

for their determination. In: Doerr and Hallauer, 1 ( Ref . 187) [71, 
72, 78] 

204a. Ellermann, V., and Bang, O. ( 1908 ) Experimentelle Leukamie bci 
Huhnern. Zentr. Bakt., Parasit., I, Orig. 46: 595-609 [324] 

205. Ellis, E. L., and Delbriick, M. (1939) The growth of bacteriophage. /. 

Gen. Physiol 22: 365-384 [41, 48, 49, 159, 162] 

206. Ellis, E. L., and Spizizen, J. (1941) The rate of bacteriophage inactiva- 

tion by filtrates of Escherichia coli cultures. J. Gen. Physiol. 24: 437-445 

207. Enders, J. F., and Pearson, H. E. (1941) Resistance of chicks to infection 

with influenza A virus. Proc. Soc. Exp. Biol Med. 48: 143-146 [227] 

208. Enders, J. F., Cohen, S., and Kano, L. W. (1945) Immunity in mumps. 

II. J. Exp. Med. 81: 119-135 [134] 

209. Ephrussi, B., and Hottinguer, H. ( 1951) On an unstable cell state in yeast. 

Cold Spring Harbor Symp. Quant. Biol. 16: 75-85 [357] 

210. Ephrussi-Taylor, H. (1951) Genetic aspects of transformations of pneu- 

mococci. Cold Spring Harbor Symp. Quant. Biol 16:445-456 [115, 

210a. Epstein, H. T. (1953) Identification of radiosensitive volume with 
nucleic acid volume. Nature 171: 384-395 [152] 

211. Esau, K. (1938) Some anatomical aspects of plant virus disease problems. 

Bot. Rev. 4: 579; 14: 413-449 [214] 

212. Fazekas de St. Groth, S., and Donnelley, M. (1950) Studies in experi- 

mental immunology of influenza. IV. Austral J. Exp. Biol Med. Sci. 
28: 61-75 [132] 

213. Feemster, R. F., and Wells, W. F. (1933) Experimental and statistical 

evidence of the particulate nature of the bacteriophage. J. Exp. Med. 

58: 385-391 [52] 
213a. Felix, A., and Anderson, E. S. (1951) Bacteriophages, virulence and 

agglutination tests with a strain of Salmonella typhi of low virulence. 

/. Hyg. 49: 349-364 [206] 
213b. Felix, A., and Callow, B. R. (1951) Paratyphoid-B Vi-phage typing. 

Lancet (2) 261: 10-14 [207] 

214. Feller, A. E., Enders, J. F., and Weller, T. H. (1940) The prolonged 

Bibliography. 387 

coexistence of vaccinia virus in high titer and living cells in roller tube 
cultures of chick embryonic tissues. J. Exp. Meet. 72: 367-388 [228] 

215. Fenner, F. (1949) Mouse pox (infectious ectromelia of mice): A review. 

J. Immunol. 63: 341-373 [64, 220] 

215a. Fenner, F., Day, M. F., and Woodroofe, G. M. (1952) The mechanism 
of the transmission of myxomatosis in the European rabbit (Oryctolagus 
cuniculus) by the mosquito Aedes aegypti. Austral. J. Exp. Biol. Med. Sci. 
30: 139-152 [320] 

216. Findlay, G. M. (1930) Warts. In: A System of Bacteriology in Relation 

to Medicine, Medical Research Council, Great Britain, 7: 252-258 [325] 

217. Findlay, G. M. (1938) Inclusion bodies and their relationships to viruses. 

In: Doerr and Hallauer, 1 (Ref. 187) [6, 230] 

218. Findlay, G. M. (1939) Variation in viruses. In: Doerr and Hallauer, 2 

(Ref. 187) [290, 298] 

219. Findlay, G. M., and Howard, E. M. (1950) Virus exaltation. Br. J. Exp. 

Path. 31: 45-50 [284] 

220. Findlay, G. M., and Ludford, R. J. (1926) The ultra-microscopic viruses. 

Br. J. Exp. Path. 7: 223-255 [230] 

221. Findlay, G. M., and MacCallum, F. O. (1937) An interference phenome- 

non in relation to yellow fever and other viruses. J. Path. Bact. 44: 405- 
424 [283] 

222. Fischer, B. (1906) Die experimented Erzeugung atypischer Epitelwuche- 

rungen und die Entstehung bosartiger Geschwiilste. Munch, med. 
Wochenschr. 53: 2041-2047 [331] 

223. Fisher, J. C., and Hollomon, J. H. (1950) A mechanism for the origin of 

cancer foci. Science 112: 424 [239] 

224. Flewett, T. II. (1953) Growth of viruses of the influenza group as re- 

vealed by electron microscopy of infected cells. In: The Nature of Virus 
Multiplication y Cambridge University Press [267] 

225. Flewett, T. H., and Challice, C. E. (1951) The intracellular growth of 

fowl-plague virus. A phase-contrast and electron microscopical study of 
infected tissue cultures. J. Gen. Microbiol. 5: 279-286 [234] 

226. Florman, A. L., and Enders, J. F. (1942) The effect of homologous anti- 

serum and complement on the multiplication of \ aceinia virus in roller- 
tube cultures of blood-mononuclear cells. /. Immunol. 43: 159-174 

227. Foster, R. A. C. ( 1948 ) The action of proflavinc on bacteriophage growth. 

J. Bact. 56: 795-809 [170, 293] 

228. Foulds, L. ( 1934 ) Filtrable tumors of fowls. Imperial Cancer Res. Fund, 

llth Rept., Suppl. [324] 

229. Foulds, L. ( 1937 ) Observations on non-filterable fowl tumours. The 

production of neutralizing sera against filtrates of Rous sarcoma I by 
non-infective extracts of a sarcoma induced by 1:2:5:6 dibenzanthracene. 
Am. J. Cancer 31: 404-413 [324] 

230. Francis, T., Jr. (1947) Dissociation of hemagglutinating and antibody- 

measuring capacities of influenza virus. J. Exp. Med. 85: 1-7 [257, 

231. Francis, T., Jr. (1950) The significance of multiple immunological types 

of influenza virus. Cincinnati J. Med. 31: 97-101 [301] 

388 Bibliography 

232. Francis, T., Jr., and Kurtz, H. B. (1950) The relation of herpes virus to 

the cell nucleus. Yale J. Biol Med. 22: 579-587 [237] 

233. Francis, T., Jr., and Salk, J. E. (1942) A simplified procedure for the 

concentration and purification of influenza virus. Science 96: 499-500 

234. Fraser, D. (1951) Bursting bacteria by release of gas pressure. Nature 

167: 33-34 [170] 

235. Fredericq, P. (1950) Rapports entre colicines et bacteriophages. Bull. 

acad. roy. med. Belgique 15: 491-510 [207] 

236. Freeman, V. J. ( 1951 ) Studies on the virulence of bacteriophage-infected 

strains of Corynebacterium diphtheriae. J. Bact. 61: 675-688 [206] 

237. French, R. C., Graham, A. F., Lesley, S. M., and Van Rooyen, C. E. ( 1952) 

The contribution of phosphorus from T2r+ bacteriophage to progeny. 
7. Bact. 64: 597-607 [176, 186] 

238. French, R. C., Lesley, S. M., Graham, A. F., and Van Rooyen, C. E. ( 1951) 

Studies on the relationship between virus and host cell. III. Canad. J. 
Med. Sci. 29: 144-148 [186, 189, 274] 

239. Friedewald, W. F., and Anderson, R. S. (1941) Influence of extraneous 

protein and virus concentration on the inactivation of the rabbit papilloma 
virus by X-rays. J. Exp. Med. 74: 463-487 [148] 

240. Friedewald, W. F., and Anderson, R. S. (1943) The effect of Roentgen 

rays on cell-virus associations. Findings with virus-induced rabbit papil- 
lomas and fibromas. J. Exp. Med. 78: 285-303 [156, 328] 

241. Friedewald, W. F., and Kidd, J. G. (1944) The recoverability of virus 

from papillomas produced therewith in domestic rabbits. /. Exp. Med. 
79: 591-605 [327, 336] 

242. Friedewald, W. F., and Pickels, E. G. (1944) Centrifugation and ultra- 

filtration studies on allantoic fluid preparations of influenza virus. J. Exp. 
Med. 79: 301-317 [253, 264] 

243. Friedrich-Freksa, H., Melchers, G., and Schramm, G. (1946) Biologischer, 

chemischer und serologischer Vergleich zweier parallelmutanten phyto- 
pathogener Viren mit ihren Ausgangsformen. Biol. Zentr. 65: 187-222 
[102, 297] 

244. Fujinami, A., and Inamoto, K. (1914) Ober Geschwiilste bei japanischen 

Haushiihnern, insbesondere liber einen transplantablen Tumor. Z. Krebs- 
forsch. 14: 94. Quoted by Oberling (Ref. 501) [324] 

245. Fukushi, T. (1940) Further studies on the dwarf disease of rice plant. 

/. Fac. Agr. Hokkaido Imp. Univ. 45: 85-152 [317] 

246. Fulton, F. (1949) Growth cycle of influenza virus. Nature 164:189 


247. Gaddum, J. H. (1933) Methods of biological assay depending on a 

quantal response. Med. Res. Counc. Spec. Rep. No. 183 [53] 

248. Card, S. ( 1943 ) Purification of poliomyelitis viruses. Experiments on 

murine and human strains. Acta Medica Scand., Suppl. 143 [45, 53, 
82, 221] 

249. Card, S., and Von Magnus, P. (1948) Studies on interference in experi- 

mental influenza. II. Ark. Kemi, Miner. Geol 24B: 1-4 [265] 

250. Garen, A., and Puck, T. T. ( 1951) The first two steps of the invasion of 

host cells by bacterial viruses. II. /. Exp. Med. 94: 177-189 [165, 
167, 260, 303] 

Bibliography 389 

250a. Gerstel, D. U. (1943) Inheritance in Nicotiana tabacum. XVII. 
Genetics 28: 533-536 [304] 

250b. Gerstel, D. U. (1948) Transfer of the mosaic-resistance factor between 
H chromosomes of Nicotiana glutinosa and N. tabacum. J. Agr. Res. 
76: 219-223 [304] 

250c. Giddings, N. J. (1950) Some interrelationships of virus strains in sugar- 
beet curly top. Phytopath. 40: 377-388 [278] 

251. Gilpatrick, J. D., and Weintraub, M. (1952) An unusual type of pro- 

tection with the carnation mosaic virus. Science 115: 701-702 [276, 

252. Ginsberg, H. S., and Horsfall, F. L., Jr. (1949) Concurrent infection with 

influenza virus and mumps virus or pneumonia virus of mice (PVM) as 
bearing on the inhibition of virus multiplication by bacterial polysaccharides. 
/. Exp. Med. 89: 37-52 [259] 

253. Ginsberg, H. S., and Horsfall, F. L., Jr. (1949) A resistant variant of 

mumps virus. Multiplication of the variant in the presence of inhibitory 
quantities of Friedlander bacillus polysaccharide. J. Exp. Med. 90: 393- 
407 [269] 

254. Ginsberg, H. S., and Horsfall, F. L., Jr. (1951) Characteristics of the 

multiplication cycle of pneumonia virus of mice (PVM). J. Exp. Med. 
93: 151-160 [222, 269] 

255. Ginsberg, H. S., and Horsfall, F. L., Jr. (1951) Therapy of infection 

with pneumonia virus of mice (PVM). Effect of polysaccharide on the 
multiplication cycles of the virus and on the course of the viral pneu- 
monia. /. Exp. Med. 93: 161-171 [222] 

256. Ginsberg, H. S., Goebel, W. F., and Horsfall, F. L., Jr. ( 1948) The inhibi- 

tory effect of polysaccharide on mumps virus multiplication. /. Exp. Med. 
87: 383-410 [259] 

256.1. Glasstone, S. K., Laidler, K. L., and Eyring, H. ( 1941 ) The Theory of 
Rate Processes. New York, McGraw-Hill [139] 

257. Gledhill, A. W., and Andrewes, C. H. (1951) A hepatitis virus of mice. 

Br. J. Exp. Path. 32: 559-568 [146] 

257a. Goebel, W. F., and Jesaitis, M. A. ( 1953) Chemical and antiviral proper- 
ties of the somatic antigens of phase II "Sh. sonnei" and of a phage resistant 
variant, II/3, 4, 7. Ann. Inst. Pasteur 84: 66-72 [168] 

258. Goodner, K. (1941) Collodion fixation; a new immunological reaction. 

Science 94:241-242 [119] 

259. Goodpasture, E. W. ( 1933) Borreliotoses: fowl pox, molluscum contagio- 

sum, variola-vaccinia. Science 77: 119-121 [18] 

260. Goodpasture, E. W., and Anderson, K. (1940) Immunity to fowl-pox stu- 

died by means of skin grafts on chorioallantois of chick embryo. Arch. 
Path. 30: 212-225 [246] 

261. Goodpasture, E. W., and Anderson, K. (1942) Virus infection of human 

fetal membranes grafted to the chorioallantois of chick embryos. Am. /. 
Path. 18: 563-575 [246] 

26?. Goodpasture, E. W., and Woodruff, C. E. (1931) A comparison of the 
inclusion bodies of fowl-pox and molluscum contagiosum. Am. J. Path. 
7: 1-8 [231] 

390 Bibliography 

263. Gough, G. A. C., and Burnet, F. M. (1934) The chemical nature of the 

phage-inactivating agent in bacterial extracts. J. Path. Bact. 38: 301-311 

264. Gowen, J. W. ( 1941) Mutation in Drosophila, bacteria and viruses. Cold 

Spring Harbor Symp. Quant. Biol 9: 187-193 [297] 

265. Gowen, J. W. (1948) Inheritance of immunity in animals. Ann. Rev. 

Microbiol 2; 215-254 [7, 305] 

266. Graham, A. F., and McClelland, L. (1950) The uptake of radioactive 

phosphorus by influenza virus A (PR8 strain). Canad. J. Res., E, 28: 121- 
134 [249] 

267. Granoff, A., Liu, O. C., and Henle, W. (1950) A small hemagglutinating 

component in preparations of Newcastle disease virus. Proc. Soc. Exp. 
Biol. Med. 75: 684-691 [264] 

268. Gratia, A. (1934) La centrifugation des bacteriophages. C. r. soc. biol. 

117: 1228-1230 [78] 

269. Gratia, A. (1936) Des relations numeriques entre bacteries lysogenes et 

particules de bacteriophage. Ann. Inst. Pasteur 57: 652-694 [41] 

270. Gratia, A. ( 1937 ) Le phenomene du halo et la synergic des bacteriophages. 

C. r. soc. biol 126:418-421 [199] 

271. Green, R. G. ( 1935) On the nature of filterable viruses. Science 82: 443- 

445 [4, 347] 

272. Green, R. G., and Stulberg, C. S. (1946) Cell-blockade in canine dis- 

temper. Proc. Soc. Exp. Biol. Med. 61: 117-121 [288] 

273. Green, R. H., and Woolley, D. W. (1947) Inhibition by certain polysac- 

charides of hemagglutination and of multiplication of influenza virus. /. 
Exp. Med. 86:55-64 [259] 

274. Green, R. H.* Anderson, T. F., and Smadel, J. E. (1942) Morphological 

structure of the virus of vaccinia. J. Exp. Med. 75: 651-656 [59, 61] 

275. Greene, H. S. N. (1941) Heterologous transplantation of mammalian 

tumors. II. /. Exp. Med. 73: 475-486 [322] 

276. Greenstein, J. P. (1947) Biochemistry of Cancer, New York, Academic 

Press [323] 

277. Gregg, N. M. (1941) Congenital cataract following German measles in 

the mother. Trans. Ophthalm. Soc. Australia 3: 35-46 [246] 

278. Greiff, D. (1948) Biology of the rickettsiae. In: The Rickettsial Diseases 

of Man, A.A.A.S., Washington [365, 369] 

279. Gross, L. ( 1951 ) "Spontaneous" leukemia developing in C3H mice follow- 

ing inoculation, in infancy, with AK-leukemic extracts, or AK embryos. 
Proc. Soc. Exp. Biol Med. 76: 27-32 [330] 

280. Gulland, J. M. (1947) The structures of nucleic acids. Cold Spring 

Harbor Symp. Quant. Biol 12: 95-103 [97] 

281. Guerin, M. C. (1905) Contr61e de la valeur des vaccins Jenneriens par la 

numeration des elements virulents. Ann. Inst. Pasteur 19: 317-320 

281a. Gupta, B. M., and Price, W. C. (1952) Mechanism of inhibition of plant 
virus infection by fungal growth products. Phytopath. 42: 45-51 [137] 

282. Gye, G. O. (1933) On improved technic for the massive tissue culture. 

Am. J. Cancer 17: 752-756 [225] 

283. Gye, G. O., and Bang, F. B. (1939) Experimental studies on the cul- 

Bibliography 391 

tural behavior and the infectivity of lymphopatia venerea virus maintained 
in tissue culture. Bull. Johns Hopkins Hosp. 65: 393-417 [229] 

284. Gye, W. E. ( 1925) The aetiology of malignant new growths. Lancet (2) 

209: 109-117 [336] 

285. Gye, W. E. (1936) Tumors transmissible by viruses. II Intern. Cancer 

Congr. 1: 48-58 [324] 

286. Hagemann, P. K. H. (1937) Fluoreszenzmikroskopische Untersuchungen 

iiber Virus und andere Mikroben. Zentr. Bakt., Parasit., I, Orig. 140: 184- 
187 [60] 

287. Hallauer, C. (1938) Die Zuchtung der Virusarten ausserhalb ihrer Wirte. 

In: Doerr and Hallauer, 1 (Ref. 187) [224, 280] 

288. Hallauer, C. (1939) Die erworbene Immunitat gegen Virusinfektionen. 

In: Doerr and Hallauer, 2 (Ref. 187) [229] 

289. Hamburger, V., and Habel, K. (1947) Teratogenetic and lethal effects of 

influenza A and mumps viruses on early chick embryos. Proc. Soc. Exp. 
Biol Med. 66: 608-617 [246, 266] 

290. Hanig, M., and Bernkopf, H. (1950) The sedimentable components of in- 

fluenza virus propagated in deembryonated eggs. /. Immune/. 65: 585-590 
[265, 266] 

291. Hardy, P. H., Jr., and Horsfall, F. L., Jr. (1948) Reactions between in- 

fluenza virus and a component of allantoic fluid. /. Exp. Med. 88: 463-484 

292. Hargett, M. V., Burruss, H. W., and Donovan, A. (1943) Aqueous-base 

yellow fever vaccine. 17. S. Publ Health Rep. 58: 505-512 [248] 

293. Haurowitz, F. (1949) Biological problems and immunochemistry. Quart. 

Rev. Biol 24: 93-101 [100] 

294. Haurowitz, F. (1950) Chemistry and Biology of Proteins, New York, 

Academic Press [98] 

295. Hayes, W. ( 1953 ) Observations on a transmissible agent determining 

sexual differentiation in Bacterium coli. J. Gen. Microbiol. 8: 72-88 

296. Heinmets, F. (1948) Studies with the electron microscope on the inter- 

action of red cells and influenza virus. /. Bact. 55: 823-831 [250, 

297. Henderson, R. G., and Topping, N. H. (1945) Epidemic typhus fever: 

Neutralization of the toxic substance. Nat. Inst. Health Bull. 183: 41-56 

298. Henle, G., and Henle, W. (1946) Studies on the toxicity of influenza 

viruses. I, II. J. Exp. Med. 84: 623-660 [31, 223, 266, 271] 

299. Henle, W. ( 1949 ) Studies on host-virus interactions in the chick embryo- 

influenza virus system. I, II. J. Exp. Med. 90: 1-22 [261, 263] 

300. Henle, W. (1950) Interference phenomena between animal viruses: A 

review. /. Immunol. 64: 203-236 [280, 284, 285, 286, 287] 

301. Henle, W., and Henle, G. (1947) The effect of ultraviolet irradiation on 

various properties of influenza viruses. 7. Exp. Med. 85: 347-364 [137, 

302. Henle, W., and Henle, G. (1949) Studies on host-virus interactions in the 

chick embryo-influenza virus system. III. /. Exp. Med. 90: 23-37 
[131, 264] 

392 Bibliography 

303. Henle, W., and Liu, O. C. (1951) Studies on host-virus interactions in 

the chick embryo-influenza virus system. VI.* 7. Exp. Med. 94: 305-322 
[156, 268] 

304. Henle, W., and Rosenberg, E. B. ( 1949) One-step growth curves of various 

strains of influenza A and B viruses and their inhibition by inactivated virus 
of the homologous type. /. Exp. Med. 89: 279-285 [261] 

305. Henle, W., Henle, G., and Rosenberg, E. B. (1947) The demonstration of 

one-step growth curves of influenza viruses through the blocking effect of 
irradiated virus on further infection. 7. Exp. Med. 86: 423-437 [261, 

306. d'Herelle, F. (1917) Sur im microbe invisible antagoniste des bacilles 

dysenteriques. C. r. acad. sci. 165: 373-375 [7] 

307. d'Herelle, F. (1921) Le bacteriophage; son role dans rimmunite. Paris, 

Masson [48] 

307a. Herriott, R. M. (1951) Nucleic acid synthesis in mustard gas-treated 
E. coli B. 7. Gen. Physiol 34: 761-764 [182] 

308. Herriott, R, M. ( 1951 ) Nucleic-acid-free T2 virus "ghosts" with specific 

biological action. J. Bact. 61: 752-754 [143, 174, 199] 

309. Hershey, A. D. (1946) Mutations of bacteriophage with respect to type 

of plaque. Genetics 31: 620-640 [184, 274, 292, 293] 

310. Hershey, A. D. (1946) Spontaneous mutations in bacterial viruses. Cold 

Spring Harbor Symp. Quant. Biol. 11: 67-76 [7, 189, 294] 

311. Hershey, A. D., and Bronfenbrenner, J. (1947) Effects of complement 

on the specific neutralization of bacteriophage. Fed. Proc. 6: 428 [126] 

312. Hershey, A. D., and Chase, M. (1951) Genetic recombination and hetero- 

zygosis in^bacteriophages. Cold Spring Harbor Symp. Quant. Biol. 16: 
471-479 [193] 

313. Hershey, A. D., and Chase, M. (1952) Independent function of viral 

protein and nucleic acid in growth of bacteriophage. 7. Gen. Physiol. 
36: 39-56 [65, 84, 104, 173, 174, 176, 337] 

314. Hershey, A. D., and Rotman, R. (1948) Linkage among genes controll- 

ing inhibition of lysis in a bacterial virus. Proc. Nat. Acad. Sci. 34: 89-96 
[189, 190, 267, 274, 292] 

315. Hershey, A. D., and Rotman, R. (1949) Genetic recombination between 

host-range and plaque-type mutants of bacteriophage in single bacterial 
cells. Genetics 34: 44-71 [189,190] 

315a. Hershey, A. D., Dixon, J., and Chase, M. (1953) Nucleic acid economy 
in bacteria infected with bacteriophage T2. I. 7. Gen. Physiol. 36: 777- 
789 [174, 177] 

316. Hershey, A. D., Kalmanson, G., and Bronfenbrenner, J. ( 1943) Quantitative 

methods in the study of the phage-antiphage reaction. 7. Immunol. 
46: 267-279 [41] 

317. Hershey, A. D., Kalmanson, G., and Bronfenbrenner, J. ( 1943) Quantitative 

relationships in the phage-antiphage reaction: unity and homogeneity of 
the reactants. 7. Immunol 46: 281-299 [125] 

318. Hershey, A. D., Kimura, F., and Bronfenbrenner, J. (1947) Uniformity 

of size of bacteriophage particles. Proc. Soc. Exp. Biol. Med. 64: 7-12 

Bibliography 393 

319. Hershey, A. D,, Kamen, M. D., Kennedy, J. W., and Gest, H. ( 1951) The 

mortality of bacteriophage containing assimilated radioactive phosphorus. 
J. Gen. Physiol 34: 305-319 [105, 154] 

320. Herzberg, K, (1927) Eine Methode zur Zahlung von Herpes-und Vak- 

zinekeimen. Zentr. Bakt., Parasit., I, Orig. 105: 57-86 [42] 

321. Herzberg, K. (1936) Der Vorgang der Vakzinevirusvermehrung in der 

Zelle. Zentr. Bakt., Parasit., I, Orig. 136: 257-260 [233, 247] 

322. Himmelweit, F. (1938) Observations on living vaccinia and ectromelia 

viruses by high power microscopy. Br. J. Exp. Path. 19: 108-123 [233, 

323. Hirst, G. K. ( 1941 ) The agglutination of red cells by allantoic fluid of chick 

embryos infected with influenza virus. Science 94: 22-23 [250] 

324. Hirst, G. K. (1942) Adsorption of influenza hemagglutinins and virus by 

red blood cells. /. Exp. Med. 76: 195-209 [250, 253] 

325. Hirst, G. K. (1943) Adsorption of influenza virus on cells of the respira- 

tory tract. /. Exp. Med. 78: 99-109 [254] 

326. Hirst, G. K. (1947) Studies on the mechanism of adaptation of influenza 

virus to mice. J. Exp. Med. 86: 357-366 [270, 301] 

327. Hirst, G. K. (1948) The nature of the virus receptors of red cells. II. 

J. Exp. Med. 87: 315-328 [142, 256] 

328. Hirst, G. K. (1949) The nature of the virus receptors of red cells. IV. 

J. Exp. Med. 89: 233-243 [257] 

329. Hirst, G. K. (1949) Laboratory diagnostic procedures for influenza. In: 

Horsfall, Diagnosis of Viral and Rickettsial Infections, New York, Columbia 
University Press [245, 249] 

330. Hirst, G. K. (1950) Receptor destruction by viruses of the mumps-NDV- 

influenza group. /. Exp. Med. 91: 161-175 [110, 254] 

331. Hirst, G. K. (1950) The relationship of the receptors of a new strain of 

virus to those of the mumps-NDV-influenza group. 7. Exp. Med. 91: 177- 
184 [254] 

332. Hirst, G. K. (1950) Hemagglutination as applied to the study of virus 

infection. In: Delbruck, Viruses 1950, Pasadena, Cal. Inst. Technol. 
Bookstore [254, 259] 

333. Holmes, F. O. (1929) Local lesions in tobacco mosaic. Bot. Gaz. 87: 39-55 

[22, 46] 

334. Holmes, F. O. ( 1930) Local and systemic increase of tobacco mosaic virus. 

Am. /. Bot. 17: 789-805 [209] 

335. Holmes, F. O. (1931) Local lesions of mosaic in Nicotiana tabacum L. 

Contrib. Boyce Thompson Inst. 3: 163-172 [24, 216] 

336. Holmes, F. O. ( 1932 ) Movement of mosaic virus from primary lesions 

in Nicotiana tabacum L. Contrib. Boyce Thompson Inst. 4: 297-322 

337. Holmes, F. O. (1934) Inheritance of ability to localize tobacco-mosaic 

virus. Phytopath. 24: 984-1002 [22, 304] 

338. Holmes, F. O. (1934) A masked strain of tobacco-mosaic virus. Phyto- 

path. 24: 845-873 [31] 

339. Holmes, F. O. (1937) Inheritance of resistance to tobacco-mosaic disease 

in the pepper. Phytopath. 27: 637-642 [22, 304] 

394 Bibliography 

339a. Holmes, F. O. (1938) Inheritance of resistance to tobacco-mosaic dis- 
ease in tobacco. Phytopath. 28: 553-561 [304] 

340. Holmes, F. O. ( 1946 ) A comparison of the experimental host ranges of 

tobacco-etch and tobacco mosaic viruses. Phytopath. 36: 643-659 [217] 

341. Holmes, F. O. (1948) Order Virales: The filterable viruses. In: Bergey's 

Manual of Determinative Bacteriology, 6th Ed., Baltimore, Williams and 
Wilkins [17] 

341a. Holmes, F. O. ( 1951 ) Indications of a new-world origin of tobacco- 
mosaic virus. Phytopath. 41: 341-349 [319] 

342. Holt, R. L,, and Kintner, J. H. (1931) Location of dengue virus in the 

body of mosquitoes. Am. J. Trop. Med. 11: 103-111 [311] 

343. Horowitz, N. H. (1945) On the evolution of biochemical syntheses. Proc. 

Nat. Acad. Set. 31: 153-157 [346] 

344. Horsfall, F. L., Jr. (1940) A low temperature storage cabinet for the 

preservation of viruses. J. Bact. 40: 559-568 [55] 

345. Horsfall, F. L., Jr. (1949) Diagnosis of Viral and Rickettsial Infections, 

New York, Columbia University Press [133] 

346. Horsfall, F. L., Jr. (1950) Approaches to the control of viral diseases. 

Bact. Rev. 14: 219-224 [146] 

347. Horsfall, F. L., Jr., and Ginsberg, H. S. ( 1951 ) The dependence of the 

pathological lesion upon the multiplication of pneumonia virus of mice 
(PVM). Kinetic relation between the degree of viral multiplication and 
the extent of pneumonia. /. Exp. Med. 93: 139-150 [222] 

348. Horsfall, F. L., Jr., and Hahn, R. G. (1940) A latent virus in normal 

mice capable of producing pneumonia in its natural host. /. Exp. Med. 
71: 391-408 [37] 

349. Horsfall, F. ., Jr., Hardy, P. H., Jr., and Davenport, F. M. (1948) The 

significance of combinations between viruses and host cells. Bull. N. Y. 
Acad. Sci. 24: 470-475 [260] 

350. Hoskins, M. ( 1935) A protective action of neurotropic against viscero- 

tropic yellow fever virus in Macacus rhesus. Am. J. Trop. Med. 15: 675- 

680 [283] 
350a. Hotchkiss, R. D. ( 1951 ) Transfer of penicillin resistance in pneumococci 

by the desoxyribonucleate derived from resistant cultures. Cold Spring 

Harbor Symp. Quant. Biol 16: 457-461 [351] 
350b. Houston, B. R., Esau, K., and Hewitt, W. B. ( 1947) The mode of vector 

feeding and the tissues involved in the transmission of Pierce's disease 

virus in grape and alfalfa. Phytopath. 37: 247-253 [317] 

351. Howe, H. A. (1952) Poliomyelitis. In: Rivers, Viral and Rickettsial Dis- 

eases of Man, 2nd Ed., Philadelphia, Lippincott [310] 

352. Howe, H. A., and Bodian, D, (1942) Neural mechanisms in poliomyelitis. 

Commonwealth Fund, New York [308] 

353. Hoyle, L. ( 1948 ) The growth cycle of influenza virus A. A study of the 

relations between virus soluble antigen and host cell in fertile eggs inocu- 
lated with influenza virus. Br. J. Exp. Path. 29: 390-399 [261, 263] 

354. Hoyle, L. ( 1950 ) The multiplication of influenza viruses in the fertile egg. 

J. Ht/g. 48: 277-297 [266] 

355. Hoyle, L., and Fairbrother, R. W. (1937) Antigenic structure of influenza 

viruses; the preparation of elementary body suspensions and the nature 
of the complement-fixing antigen. J. Ht/g. 37: 512-520 [130, 131] 

Bibliography 395 

355a. Hughues, F. M. ( 1952) Development of the inclusion bodies of a granu- 
losis virus. J. Bact. 64: 375-380 [84] 

356. Hurlbut, H. S. (1950) The recovery of poliomyelitis virus after parenteral 

introduction into cockroaches and houseflies. /. Infectious Diseases 86: 
103-104 [309] 

356a. Hutton, E. M., and Wark, D. C. ( 1952) A relationship between immunity 
and localized reaction to virus X in the potato (Solanum tuberosum). 
Austral J. Sci. Res. B, 5: 237-243 [24] 

357. Hyden, H. (1947) The nucleoproteins in virus reproduction. Cold Spring 

Harbor Symp. Quant. Biol. 12: 104-114 [60, 238] 

358. Isaacs, A., and Edney, M. (1950) Interference between inactive and active 

viruses in the chick embryo. I. Austral J. Exp. Biol Med. Sci. 28: 219- 
230 [284] 

359. Iwanowski, D. (1903) Ober die Mosaikkrankheit der Tabakspflanze. 

Z. Pflanzenkr. 13: 1-41 [214] 

359a. Jacob, F. (1952) Dc\ eloppemcnt spontane et induit des bactcriophages 
chez dcs Pseudomonas pyocyanea polylysogones. Arm. Inst. Pasteur 83: 
671-692 [202] 

360. Jacob, F., Siminovitch, L., and Wollman, E. (1951) Induction dc la pro- 

duction d'une colicinc par le rayonncment ultraviolet. C. r. acad. set. 
233: 1500-1502 [207] 

361. Jcnner, E. (1798) An inquiry into the causes and effects of the variolae 

vaccinae, a disease discovered in some of the western counties of England, 
particularly Gloucestershire, and known by the name of the cow pox. 
London [6] 

361a. Jensen, J. H. (1933) Isolation of yellow-mosaic viruses from plants in- 
fected with tobacco mosaic. Phytopath. 23: 964-974 [7] 

362. Jensen, J. II. (1937) Studies on representative strains of tobacco mosaic 

virus. Phytopath. 27: 69-84 [296] 

363. Johnson, F. (1941) Transmission of plant \iruses by dodder. Phytopath. 

31: 649-656 [34] 

364. Kalmanson, G., and Bronfcnbrcnncr, S. (1939) Studies on the purification 

of bacteriophage. /. Gen. Physiol 23: 203-228 [103] 

365. Kalmanson, G. M., and Bronfenbrcnner, J. ( 1942) Evidence of serological 

heterogeneity of polyvalent "pure line" bacteriophage. /. Immunol 45: 
13-19 [123] 

366. Kalmanson, G. M., and Brontenbrenner, J. (1943) Restoration of activity 

of neutralized biologic agents by removal of the antibody with papain. 
J. Immunol. 47: 387-407 [123, 125] 

367. Kassanis, B. (1939) Intranuclear inclusions in virus-infected plants. 

Ann. Appl Biol. 26: 705-709 [215] 

368. Kausche, G. A. (1940) t)ber eine das Virusprotein inaktivierende Sub- 

stanz in Samen von Nicotiana tabacum Samsun. Biol Zentr. 60: 423-438 

369. Keogh, E. V. (1938) Ectodermal lesions produced by the virus of Rons 

sarcoma. Br. J. Exp. Path. 19: 1-9 [246] 

370. Kidd, J. G., and Rons, P. (1938) The carcinogenic effect of a papilloma 

virus on the tarred skin of rabbits. II. J. Exp. Med. 68: 529-562 [338] 

371. Kidd, J. G. (1942) The enduring partnership of a neoplastic virus and 

carcinoma cells. /. Exp. Med. 75: 7-20 [327] 

396 Bibliography 

372. Kidd, J. G. ( 1946 ) Suppression of growth of Brown-Pearce tumor cells by 

a specific antibody. 7. Exp. Med. 83: 227-250 [337] 

373. Kidd, J. G., and Rous, P. (1940) Cancers deriving from the virus papil- 

lomas of wild rabbits under natural conditions. /. Exp. Med. 71: 469-494 

374. Kirber, H. P., Kirber, M. W., and Henle, W. (1950) Transplantation of 

cornea to chorioallantoic membrane, with observations on virus inocula- 
tion. Proc. Soc. Exp. Biol. Med. 73: 481-485 [246] 

375. Kleczkowski, A. (1950) Interpreting relationships between the concentra- 

tions of plant virus and numbers of local lesions. /. Gen. Microbiol. 4: 
53-69 [54] 

376. Kligler, I. J., and Olenick, E. (1943) Inactivation of phage by aldehydes 

and aldoses and subsequent reactivation. J. Immunol. 47: 325-333 

377. Knight, C. A. (1946) Precipitin reactions of highly purified influenza 

viruses and related materials. J. Exp. Med. 83: 281-294 [107, 260, 

378. Knight, C. A. (1947) The nucleic acid and carbohydrate of influenza 

virus. J. Exp. Med. 85: 99-116 [107] 

379. Knight, C. A. (1947) Amino acid composition of highly purified viral 

particles of influenza A and B. 7. Exp. Med. 86: 125-129 [108] 

380. Knight, C. A. (1947) The nature of some of the chemical differences 

among strains of tobacco mosaic virus. J. Biol. Chem. 171: 297-308 
[102, 297] 

381. Knight, C. A. (1949) Constituents of viruses. Ann. Rev. Microbiol. 

3: 121-136 [103] 

382. Koch, A. (1936) Symbiosestudien. II. Experimentelle Untersuchungen 

an Oryzaephilus surinamensis L. (Cucujidae, Coleopt. ). Z Morphol. 
Okol. Tiere 32: 137-180 [370] 

383. Koprowski, H. (1950) Immunological reactions in viral diseases. Ann. 

Rev. Microbiol. 4:261-282 [111, 179] 

383a. Kozloff, L. M. (1953) Desoxyribonuclease inhibitor and bacteriophage 
infection. Fed. Proc. 12:234 [179] 

384. Kozloff, L., Putnam, F. W., and Evans, E. A., Jr. (1950) Precursors of 

bacteriophage nitrogen and carbon. In: Viruses 1 950, Pasadena, Cal. 
Inst. Technol. Bookstore [176] 

385. Krueger, A. P. ( 1937 ) The mechanism of bacteriophage production. 

Science 86: 379-380 [178, 345] 

386. Kunkel, L. O. (1934) Studies on acquired immunity with tobacco and 

aucuba mosaics. Phytopath. 24: 437-466 [277] 

387. Kunkel, L. O. ( 1936 ) Immunological studies on the three peach diseases, 

yellows, rosette, and little peach. Phytopath. 26: 201-219 [278] 

388. Kunkel, L. O. (1936) Heat treatments for the cure of yellows and other 

virus diseases of peach. Phytopath. 26: 809-830 [31] 

389. Kunkel, L. O. (1937) Effect of heat on ability of Cicadula sexnotata 

(Fall.) to transmit aster yellows. Am. J. Bot. 24: 312-327 [138, 317] 

390. Kunkel, L. O. (1941) Heat cure of aster yellows in periwinkles. Am. 7. 

Bot. 28:761-769 [138, 145, 317] 

Bibliography 397 

391. Kunkel, L. O. (1943) New hosts as a key to progress in plant virus 

disease research. In: Virus Diseases, Ithaca, Cornell University Press 
[34, 36] 

392. Labaw, L. W. (1951) The origin of phosphorus in Escherichia coli bac- 

teriophages. /. Bad. 62: 169-173 [176] 

393. Lacassagne, A. ( 1936 ) . Hormonal pathogenesis of adcnocarcinoma of the 

breast. Am. J. Cancer 27: 217-228 [329] 

394. Lahelle, O., and Horsfall, F. L., Jr. ( 1949) Multiplication of influenza virus 

in dead chick embryos. Proc. Soc. Exp. Biol. Med. 70: 547-551 [245] 

395. Laidlaw, P. P. (1938) Virus Diseases and Viruses, Cambridge, Cambridge 

University Press [4, 347] 

396. Laidlaw, P. P., and Elford, W. J. ( 1936) A new group of filterable organ- 

isms. Proc. Roy. Soc. (London) B, 120: 292-303 [4] 

397. Landsteiner, K. (1945) The Specificity of Serological Reactions, Cam- 

bridge, Harvard University Press [117] 

398. Lanni, F., and Beard, J. W. ( 1948) Inhibition by egg-white of hemaggluti- 

nation by swine influenza virus. Proc. Soc. Exp. Biol. Med. 68: 312-313 

399. Lanni, F., and Lanni, Y. T. (1952) A quantitative theory of influenza 

virus hemagglutination-inhibition. /. Bact. 64: 865-882 [258] 
399a. Lanni, F., and Lanni, Y. T. (1953) Antigenic structure of bacteriophage 
T2. Fed. Proc. 12: 450-451 [126, 127] 

400. Lanni, F., Sharp, D. C., Eckert, E. A., Dillon, E. S., Beard, D., and Beard, 

J. W. ( 1949 ) The egg white inhibitor of influenza virus hemagglutination. 
I. /. Biol. Chem. 179: 1275-1287 [258] 

401. Lanni, Y. T. (1953) Infection by bacteriophage T5 and the intracellular 

formation of phage antigen and active phage. Bact. Proc. 39 [174] 

402. Latarjet, R., and Wahl, R. ( 1945) Precisions sur 1'inactivation des bacterio- 

phages par les rayons ultraviolets. Ann. Inst. Pasteur 71: 336-339 [154] 

403. Lauffer, M. A. (1943) The -sedimentation rate of the infectious principle 

of tobacco mosaic virus. /. Biol Chem. 151: 627-634 [93] 

404. Lauffer, M. A., and Price, W. C. (1945) Infection by viruses. Arch. 

Biochem. 8: 449-468 [50, 51, 54] 

405. Lauffer, M. A., Price, W. C., and Petre, A. W. (1949) The nature of 

viruses. Adv. Enzymol 9: 171-240 [7, 79, 82, 93, 96, 140, 141, 144] 

406. Law, L. W. ( 1942 ) Foster nursing and the growth of transplantable 

leukemias in mice. Cancer Res. 2: 108-115 [322] 

407. Lea, D. E. ( 1946) Actions of Radiations on Living Cells, Cambridge Uni- 

versity Press [150, 151, 152] 

408. Lea, D., Smith, K. M., Holmes, B., and Markham, R, (1944) Direct and 

indirect actions of radiations on viruses and enzymes. Parasitology 36: 110- 
118 [148] 

409. Lederberg, E. M., and Lederberg, J. (1953) Genetic studies of lysogenic- 

ity in Escherichia coli. Genetics 38: 51-64 [205, 303] 

410. Lederberg, J. (1949) Aberrant heterozygotes in Escherichia coli. Proc. 

Nat. Acad. Set. 35: 178-184 [303] 

411. Lederberg, J., and Lederberg, E. M. (1952) Replica plating and indirect 

selection of bacterial mutants. /. Bact. 63: 399-406 [303] 

412. Lederberg, J., and Tatum, E. L. (1946) Novel genotypes in mixed cul- 

398 Bibliography 

tures of biochemical mutants of bacteria. Cold Spring Harbor Symp. 
Quant. Biol. 11: 113-114 [205] 

413. Lederberg, J., Cavalli, L. L., and Lederberg, E. M. (1952) Sex compati- 

bility in Escherichia coli. Genetics 37: 720-730 [353] 

414. Levaditi, C., Lepine, P., and Verge, J. (1943) Les ultravirus des maladies 

animates, Paris, Maloine [29] 

415. Levine, P., and Frisch, A. W. (1934) On specific inhibition of bacterio- 

phage action by bacterial extracts. /. Exp. Med. 59: 213-228 [168] 

416. Levinson, S. O., Milzer, A., Shaughnessy, H. J., Neal, J. L., and Oppen- 

heimer, F. ( 1944 ) Production of potent inactivated vaccines with ultra- 
violet irradiation. II. /. Am. Med. Assoc. 125: 531-532 [154] 
416a. Levinthal, C. (1953) Recombination in phage: its relationship to hetero- 
zygosis and growth. Genetics (in press) [194] 

417. Levinthal, C., and Fisher, H. (1952) The structural development of a 

bacterial virus. Biochem. Biophys. Acta 9: 419-429 [172, 175] 

418. Levinthal, W. (1935) Recent observations on psittacosis. Lancet (1) 

228: 1207-1210 [232, 247] 

419. Lewis, D. J., Hughes, T. P., and Mahaffy, A. F. ( 1942) Experimental trans- 

mission of yellow fever by three common species of mosquitoes from the 
Anglo-Egyptian Sudan. Ann. Trop. Med. Parasit. 36: 34-37 [311] 

420. Lewis, E. B. (1951) Pseudoallelism and gene evolution. Cold Spring 

Harbor Symp. Quant. Biol. 16: 159-174 [360] 

421. L'Hritier, P. (1951) The CO 2 sensitivity problem in Drosophila. Cold 

Spring Harbor Symp. Quant. Biol. 16: 99-112 [357] 

422. L'Heritier, P., and Plus, N. (1950) Inactivation par les rayons X du 

virus resppnsable de la sensibilite au CO 2 chez la Drosophile. C. r. acad. 
sci. 231: 192-194 [357] 

423. Li, C. P., and Rivers, T. M. (1930) Cultivation of vaccine virus. /. Exp. 

Med. 52: 465-470 [225] 

424. Lieb, M. (1953) The establishment of lysogenicity in Escherichia coli. 

J. Bact. 65: 642-651 [203] 

425. Little, C. C. (1947) The genetics of cancer in mice. Biol. Rev. 22: 315- 

343 [322, 339] 

426. Liu, O. C., and Henle, ,W. (1951) Studies on host-virus interactions in 

the chick embryo-influenza virus system. IV. J. Exp. Med. 94: 269-289 
[263, 267, 287] 

427. Liu, O. C., and Henle, W. (1951) Studies on host-virus interactions in 

the chick embryo-influenza virus system. V. /. Exp. Med. 94: 291-304 

428. Loeffler, F., and Frosch, P. (1898) Berichte der Kommission zur Erfor- 

schung der Maul- und Klauenseuche bei dem Institut fur Infektionskrank- 
heiten in Berlin. Zentr. Bakt., Parasit., I, Orig. 23: 371-391 [6] 

429. Lucke, B. (1938) Carcinoma in the leopard frog: its probable causation 

by a virus. /. Exp. Med. 68: 457-468 [325, 333] 

430. Ludford, R. J. ( 1937 ) The production of tumours by cultures of normal cells 

treated with filtrates of filterable fowl tumours. Am. J. Cancer 31: 414- 
429 [332] 

431. Luria, S. (1940) Methodes statistiques appliquees aTetude du mode d'ac- 

tion des ultravirus. Ann. Inst. Pasteur 64: 415-436 [48, 52] 

Bibliography 399 

432. Luria, S. E. (1944) Dissociation of the growth of bacterial viruses and of 

their host by means of temperatures above optimum. Proc. Indiana Acad 
Sci. 53: 28-29 [138] 

433. Luria, S. E. (1944) A growth-delaying effect of ultraviolet radiation on 

bacterial viruses. Proc. Nat. Acad. Sci. 30: 393-397 [137, 153] 

434. Luria, S. E. (1945) Mutation of bacterial viruses affecting their host 

range. Genetics 30: 84-99 [183, 184, 293] 

435. Luria, S. E. (1946) Spontaneous bacterial mutations to resistance to 

anti-bacterial agents. Cold Spring Harbor Symp. Quant. Biol 11: 130-137 
[187, 303] 

436. Luria, S. E. (1947) Reactivation of irradiated bacteriophage by transfer 

of self -reproducing units. Proc. Nat. Acad. Sci. 33: 253-264 [156, 

437. Luria, S. E. (1950) Bacteriophage: An essay on virus reproduction. 

Science 111:507-511 [182, 353] 

438. Luria, S. E. (1951) The frequency distribution of spontaneous bacterio- 

phage mutants as evidence for the exponential rate of phage reproduction. 
Cold Spring Harbor Symp. Quant. Biol. 16: 463-470 [185, 292] 

439. Luria, S. E. (1953) An analysis of bacteriophage multiplication. In: The 

Nature of Virus Multiplication, Cambridge University Press [84] 

440. Luria, S. E., and Anderson, T. F. ( 1942) The identification and characteri- 

zation of bacteriophages with the electron microscope. Proc. Nat. Acad. 
Sci. 28: 127-130 [67] 

441. Luria, S. E., and Delbriick, M. (1942) Interference between inactivated 

bacterial virus and active virus of the same strain and of a different 
strain. Arch. Biochem. 1: 207-218 [165, 168, 194] 

442. Luria, S. E., and Delbriick, M. (1943) Mutations in bacteria from virus 

sensitivity to virus resistance. Genetics 28: 491-511 [186, 303] 

443. Luria, S. E., and Dulbecco, R. (1949) Genetic recombination leading to 

production of active bacteriophage from ultraviolet inactivated bacterio- 
phage particles. Genetics 34:93-125 [195] 

444. Luria, S. E., and Exner, F. M. (1941) The inactivation of bacteriophages 

by X-rays. Influence of the medium. Proc. Nat. Acad. Sci. 27: 370-375 
[148, 150] 

445. Luria, S. E., and Human, M. L. (1950) Chromatin staining of bacteria 

during bacteriophage infection. /. Bact. 59: 551-560 [169, 180] 

446. Luria, S. E., and Human, M. L. (1952) A nonhereditary, host-induced 

variation of bacterial viruses. /. Bact. 64:557-569 [197, 207, 295, 

447. Luria, S. E., and Latarjet, R. (1947) Ultraviolet irradiation of bacterio- 

phage during intracellular growth. J. Bact. 53: 149-163 [195, 196] 

448. Luria, S. E., Delbruck, M., and Anderson, T. F. (1943) Electron micro- 

scope studies of bacterial viruses. /. Bact. 46: 57-77 [67, 69, 159] 

449. Luria, S. E., Williams, R. C., and Backus, R. C. (1951) Electron micro- 

graphic counts of bacteriophage particles. J. Bact. 61: 179-188 [53, 
54, 68] 

450. Lwoff, A. (1943) Evolution physiologique. tude des pertes de fonctions 

chez les microorganismes, Paris, Hermann [346] 

400 Bibliography 

451. Lwoff, A. ( 1950) Problems of Morphogenesis in Ciliates, New York, Wiley 


452. Lwoff, A. (1953) The nature of phage production. In: The Nature of 

Virus Multiplication, Cambridge University Press [201] 

453. Lwoff, A., and Guttman, A. (1950) Recherches sur un Bacillus mega- 

therium lysogene. Ann. Ins*. Pasteur 78: 711-738 [30, 201] 

454. Lwoff, A., Siminovitch, L., and Kjeldgaard, N. ( 1950) Induction de la pro- 

duction de bact^riophages chez une bacterie lysogene. Ann. Inst. Pas- 
teur 79: 815-859 [156, 201, 328] 

455. MacCallum, F. O. (1936) The necessity of living cells for the cultivation 

of psittacosis virus. Br. J. Exp. Path. 17: 472-481 [228] 

456. Mclntosh, J. ( 1933) On the nature of the tumors induced in fowls by in- 

jections of tar. Br. J. Exp. Path. 14: 422-434 [338] 

457. McKinney, H. H. (1926) Virus mixtures that may not be detected in 

young tobacco plants. Phytopath. 16: 893 [7, 297] 

458. McKinney, H. H. (1929) Mosaic diseases in the Canary Islands, West 

Africa, and Gilbraltar. /. Agr. Res. 39: 557-578 [276] 

459. McKinney, H. H. ( 1935 ) The inhibiting influence of a virus on one of 

its mutants. Science 82: 463-464 [278] 

460. McKinney, H. H. (1941) Virus-antagonism tests and their limitations for 

establishing relationship between mutants, and nonrelationship between 
distinct viruses. Am. J. Bot. 28: 770-778 [278] 

461. McLean, I. W., Jr., Beard, D., Taylor, A. R., Sharp, D. G., Beard, J. N., 

Feller, A. E., and Dingle, J. H. (1944) Influence of temperature of 
incubation on the increase of influenzal virus B (Lee strain) in the 
chorio-albmtoic fluid of chick embryos. /. Immunol 48: 305-316 [248] 

462. McWhorter, F. P., and Price, W. C. (1949) Evidence that two different 

plant viruses can multiply simultaneously in the same cell. Science 
109: 116-117 [215, 279] 

463. Maal0e, O., and Stent, G. S. ( 1952) Radioactive phosphorus tracer studies 

on the reproduction of T4 bacteriophage. I. Acta Path. Microb. Scand. 
30: 149-157 [176] 

464. Maal0e, O., and Symonds, N. (1953) Radioactive sulfur tracer studies 

on the reproduction of T4 bacteriophage. J. Bact. 65: 177-182 [172, 

465. Maal0e, O. C., and Watson, J. D. (1951) The transfer of radioactive 

phosphorus from parental to progeny phage. Proc. Nat. Acad. Sci. 37: 
507-513 [176] 

466. Magrassi, F. (1935) Studii sull'infezione e sull'immunita da virus erpetico. 

II. Z. Hyg. Infectionskrh. 117: 501-528 [280] 

467. Magrassi, F. (1935) Studii sull'infezione e suirimmunit da virus erpetico. 

III. Z. Hyg. Infectionskrh. 117: 573-620 [282] 

468. Maitland, H. B., and Laing, A. W. (1930) Experiments on the cultivation 

of vaccinia virus. Br. J. Exp. Path. 11: 119-126 [229] 

469. Maitland, H. B., and Maitland, M. C. (1928) Cultivation of vaccinia 

virus without tissue culture. Lancet (2) 215: 596-597 [225] 

470. Maramorosch, K. (1952) Direct evidence for the multiplication of aster- 

yellows virus in its insect vector. Phytopath. 42: 59-64 [318] 

Bibliography 401 

471. Markham, R. (1953) Nucleic acids in virus multiplication. In: The 

Nature of Virus Multiplication, Cambridge University Press [97 100 
101, 110] 

472. Markham, R., and Smith, J. D. (1950) Chromatographic studies on nucleic 

acids. Biochem. J. 46: 513-517 [100, 101] 

473. Markham, R., and Smith, K. M. (1949) Studies on the virus of turnip 

yellow mosaic. Parasitology 39: 330-342 [91, 101, 110, 128, 215] 

474. Markham, R., Smith, K. M., and Lea, D. ( 1942) The sizes of viruses and 

the methods employed in their estimation. Parasitology 34: 315-352 

475. Markham, R., Smith, K. M., and Wyckoff, R, W. G. (1948) Molecular 

arrangement in tobacco necrosis virus crystals. Nature 161: 760-761 
[91, 92] 

476. Mason, J. H., Coles, J. D. W. A., and Alexander, R. A. (1940) Cultiva- 

tion of bluetongue virus in fertile eggs produced on a vitamin deficient 
diet. Nature 145: 1022 [223] 

477. Matthews, R. E. F. (1953) Chemotherapy and plant viruses. J. Gen. 

Microbiol 8: 277-288 [146, 210] 

478. Melnick, J. L. ( 1947 ) Poliomyelitis virus in urban sewage in epidemic and 

in non-epidemic times. Am. J. Hyg. 45: 240-253 [309] 

479. Meneghini, M., and Delwiche, C. C. (1951) The multiplication of to- 

bacco mosaic virus in the host tobacco plant. /. Biol. Chem. 189: 177-186 
[210, 214] 

480. Merrill, M. H. (1936) The mass factor in immunological studies upon 

viruses. /. Immunol 30: 169-184 [118] 

481. Merrill, M. H., and TenBroeck, C. (1935) The transmission of equine 

encephalomyelitis virus by Aedes aegypti. J. Exp. Med. 62: 687-695 
[311, 312] 

482. Merrill, M. H., Lacaillade, C. W., and TenBroeck, C. (1934) Mosquito 

transmission of equine encephalomyelitis. Science 80: 251-252 [312] 

483. Meyer, K. F. (1942) The ecology of psittacosis and ornithosis. Medicine 

21: 175-206 [302] 

484. Miller, E. M., and Goebel, W. F. (1949) Studies on bacteriophage. I. 

/. Exp. Med. 90: 255-265 [168, 284] 

485. Miller, G. L., and LaufFer, M. A., and Stanley, W. M. (1944) Electropho- 

retic studies on PR8 influenza virus. J. Exp. Med. 80: 549-559 [80] 

486. Mills, K. C., and Dochez, A. R. (1944) Specific agglutination of murine 

erythrocytes by a pneumonia virus in mice. Proc. Soc. Exp. Biol. Med. 
57: 140-143 [37, 340] 

486a. Mommaerts, E. B., Eckert, E. A., Beard, D., Sharp, D. G., and Beard, J. W. 
(1952) Dephosphorylation of adenosine triphospbate by concentrates 
of the virus of avian erythromyeloblastic leucosis. Proc. Soc. Exp. Biol. 
Med. 79:450-455 [108] 

487. Monod, J. (1941) The phenomenon of enzymatic adaptation. Growth 

11: 223-289 [358] 

488. Monod, J., and Cohn, M. (1952) La biosynthese induite des enzymes 

(adaptation enzymatique ) . Adv. Enzymol 13:67-119 [358] 

489. Monod, J., and Wollman, E. (1947) L'inhibition de la croissance et de 

402 Bibliography 

1'adaptation enzymatique chez les bacteries infectees par le bacteriophage. 
Ann. Inst. Pasteur 73: 937-957 [179] 

489a. Moore, A. E. (1952) Viruses with oncolytic properties and their adapta- 
tion to tumors. Ann, N. Y. Acad. Set. 54: 945-952 [306] 

490. Morgan, C., and Wyckoff, R. W. G. (1950) The electron microscopy of 

fowl pox virus within the chorioallantoic membrane. J. Immunol. 65: 
285-295 [234] 

491. Morgan, I. M. (1941) Influence of age on susceptibility and on immune 

response of mice to Eastern equine encephalomyelitis virus. /. Exp. 
Med. 74: 115-132 [306] 

492. Morgan, I. M. (1947) The role of antibody in experimental poliomyelitis. 

III. Am. J. Hyg. 45: 390-400 [282, 314] 

493. Morgan, I. M. (1949) Mechanism of immunity in poliomyelitis and its 

bearing on differentiation of types. Am. J. Med. 6: 556-562 [133] 

494. Mosley, V. M., and Wyckoff, R. W. G. (1946) Electron micrography of 

the virus of influenza. Nature 157: 263 [82] 

495. Muckenfuss, R. S., and Rivers, T. M. (1930) Survival of vaccine virus 

separated from living host cells by collodion membranes. J. Exp. Med. 
51: 149-159 [227] 

496. Muller, H. J. (1922) Variation due to change in the individual gene. 

Am. Naturalist 56: 32-50 [350] 

496a. Muller, R. H., and Rose, H. M. (1952) Purification of influenza virus 
(PR8 strain) by a cation exchange resin. Proc. Soc. Exp. Biol. Med. 
80: 27-29 [89] 

497. Murphy, J. B. ( 1913) Transplantability of tissues to the embryo of foreign 

species. /. Exp. Med. 17: 482-493 [322] 

498. Murray, R. T. E., Gillen, D. H., and Heagy, F. G. (1950) Cytological 

changes in Escherichia coli produced by infection with phage T2. J. Bact. 
59: 603-615 [180] 

498a. Nolla, J. A. B. (1938) Inheritance in Nicotiana. III. /. Hered. 29: 42- 
48 [304] 

499. Northrop, J. H. (1939) Crystalline Enzymes, New York, Columbia Uni- 

versity Press [99, 103] 

500. Novick, A., and Szilard, L. (1951) Virus strains of identical phenotype 

but different genotype. Science 113: 34-35 [295] 

501. Oberling, C. (1952) The Riddle of Cancer, 2nd Ed., New Haven, Yale 

University Press [332, 335] 

502. Oberling, C., Guerin, M., and Guerin, P. (1935) A propos de la trans- 

formation sarcomateuse des fibroadenomes mammaires transplantables du 
rat blanc. Bull, assoc. franc. &tude cancer 24: 232 [323] 

503. Ocfemia, G. O. (1934) An insect vector of the Fiji disease of sugar cane. 

Am. J. Bot. 22: 113-120 [216] 

504. Oparin, A. I. ( 1938) The Origin of Life, New York, Macmillan [346] 

505. Orlando, A., and Silberschmidt, K. (1946) Quoted by Bawden (Ref. 43) 


506. Oster, G., and McLaren, A. D. ( 1950) The ultraviolet light and photosensi- 

tized inactivation of tobacco mosaic virus. /. Gen. Physiol. 33: 215-228 

Bibliography 403 

507. Oster, G., and Stanley, W. M. (1946) An electronmicroscope study of the 

contents of hair cells from leaves diseased with tobacco mosaic virus. Br. 
J. Exp. Path. 27: 261-265 [93] 

508. Pang, K. H., and Zia, S. H. (1940) Studies of virus of encephalitis (St. 

Louis type ) grown on agar tissue medium. Chinese Med. /., Suppl. 3:446- 

457 [225] 
508a. Pardee, A. B., and Williams, I. ( 1953) Enzymatic activity and bacterio- 

phage infection. III. Ann. Inst. Pasteur 84: 147-156 [179] 
508b. Parker, R. C. (1950) Methods of Tissue Cultures, 2nd Ed., New York, 

Hoeber [225] 

509. Parker, R. F. (1938) Statistical studies of the nature of the infectious 

unit of vaccine virus. /. Exp. Med. 67: 725-738 [51, 53] 

510. Parker, R. F. (1940) Studies on the infectious unit of myxoma. /. Exp. 

Med. 71: 439-444 [53] 

511. Parker, R. F., and Rivers, T. M. (1936) Immunological and chemical 

investigations of vaccine virus. IV. /. Exp. Med. 64: 439-452 [44] 

512. Parker, R. F., Bronson, L. H., and Green, R. H. (1941) Further studies 

on the infectious unit of vaccinia. J. Exp. Med. 74: 263-281 [53] 

513. Parsons, R. J., and Kidd, J. G. (1943) Oral papillomatosis of rabbits: a 

virus disease. J. Exp. Med. 77: 233-250 [326] 

514. Paschen, E. (1917) Technik zur Darstellung der Elementarkorperchen 

(Paschensche Korperchen) in der Variolapustel. Deutsch. med. Wochen- 
schr. 43: 1036 [6, 58] 

515. Peacock, P. R. (1935) Studies of fowl tumours induced by carcinogenic 

agents. II. Am. /. Cancer 25: 49-65 [332, 338] 

516. Perdrau, J. R., and Todd, C. ( 1933) The photodynamic action of methyl- 

ene blue on certain viruses. Proc. Roy. Soc. (London) B, 112:288-298 

517. Pinkerton, H. (1942) The pathogenic rickettsiae with particular reference 

to their nature, biologic properties, and classification. Bact. Rev. 6: 37-78 

518. Pinkerton, H. (1948) The classification of rickettsiae and rickettsial dis- 

eases. In: The Rickettsial Diseases of Man, A.A.A.S., Washington 
[364, 365, 372] 

519. Pirie, A. (1937) The metabolism of the filter-passing organism "C" from 

sewage. Br. J. Exp. Path. 18: 96-102 [4] 

520. Pirie, N. W. (1940) The criteria of purity used in the study of large 

molecules of biological origin. Biol Rev. 15:377-404 [85] 

521. Pirie, N. W. (1945) Physical and chemical properties of tomato bushy 

stunt virus and the strains of tobacco mosaic virus. Adv. Enzymol. 
5: 1-29 [56] 

522. Pirie, N. W. (1950) A biochemical approach to viruses. Nature 166: 495- 

496 [86] 

523. Plotz, H., and Ephrussi, B. (1933) La culture de la peste aviaire en 

presence de cellules vivantes non proliferates. C. r. soc. biol. 113: 711- 
712 [227] 

524. Poisson, S. D. (1837) Recherches sur la probabilite des jugements en 

matiere criminelle et en matiere civile, pr6c6dees des regies generates du 
calcul des probabilites. Paris [49] 

404 Bibliography 

525. Poison, A., and Shepard, C. C. (1949) On the diffusion rates of bacterio- 

phages. Biochem. Biophys. Acta 3: 137-145 [79] 

526. Poison, A., and Wyckoff, R. W. G. (1948) The amino acid content of 

bacteriophage. Science 108:501 [104] 

527. Porter, K. R., and Thompson, A. P. (1948) A particulate body associated 

with epithelial cells cultured from mammary carcinomas of mice of a 
milk-factor strain. /. Exp. Med. 88: 15-24 [328, 329] 

528. Porter, K. R., Claude, A., and Fullam, E. F. (1945) A study of tissue 

culture cells by electron microscopy. Methods and preliminary observa- 
tions. /. Exp. Med. 81: 233-246 [234] 

529. Preer, J. R. ( 1948 ) A study of some properties of the cytoplasmic factor 

"kappa" in Paramecium aurelia, variety 2. Genetics 33: 349-404 [357] 

530. Preer, J. R. (1946) Some properties of a genetic cytoplasmic factor in 

Paramecium. Proc. Nat. Acad. Sci. 32: 247-253 [339] 

531. Price, W. C. (1934) Isolation and study of some yellow strains of cu- 

cumber mosaic. Phytopath. 24: 743-761 [297] 

532. Price, W. C. (1938) Studies on the virus of tobacco necrosis. Am. J. Bot. 

25: 603-612 [308] 

533. Price, W. C. (1940) Acquired immunity from plant virus diseases. Quart. 

Rev. Biol 15: 338-361 [276, 278] 

533a. Price, W. C. (1943) Severity of curly top in tobacco affected by site 
of inoculation. Phytopath. 33: 586-601 [279] 

534. Price, W. H. (1947) Bacteriophage formation without bacterial growth. 

I. 7. Gen. Physiol. 31: 119-126 [179] 

535. Puck, T. T., Garen, A., and Cline, J. (1951) The mechanism of virus 

attachment to host cells. I. J. Exp. Med. 93: 65-88 [165, 166, 260] 

536. Purdy, H. A^. (1929) Immunologic reactions with tobacco mosaic virus. 

/. Exp. Med. 49: 919-935 [127] 

537. Purdy, W. J. (1933) The hetero-transfer of two filterable tumours: an 

investigation by means of resistance induced by embryo tissue. Br. J. 
Exp. Path. 14: 9-17 [332] 

538. Purdy, W. J. (1933) The hetero-transfer of two filterable tumours: an 

investigation by means of immune sera. Br. J. Exp. Path. 14: 260-267 

539. Putnam, F. W. (1950) Molecular kinetics and electrophoretic properties 

of bacteriophages. Science 111: 481-488 [79, 80] 

540. Rafelson, M. E., Jr., Winzler, R. J., and Pearson, H. E. ( 1949) The effect 

of Theiler's GD VII virus on P 32 uptake by minced one-day-old mouse 
brain. J. Biol Chem. 181: 583-593 [238] 

541. Raff, N., and Cohen, S. S. (1950) The effect of virus infection on the 

utilization of tryptophan by Escherichia coli. J. Bact. 60: 69-80 [179] 

542. Raffel, S., and Schulz, E. W. (1940) Immunological reactions in polio- 

myelitis. J. Immunol. 39: 265-286 [49, 134] 

543. Rake, G. (1948) Chlamydozoaceae Moshkovsky. In: Sergey's Manual of 

Determinative Bacteriology, 6th Ed., Baltimore, Williams and Wilkins 

544. Rake, G., and Jones, H. P. ( 1944) Studies on lymphogranuloma venereum. 

I, II. J. Exp. Med. 75t 323-338; 79: 463-486 [223, 247] 

545. Rakieten, M. L., and Rakieten, T. L. ( 1937) Relationship between staphy- 

Bibliography 405 

lococci and bacilli belonging to the subtilis group as shown by bacterio- 
phage absorption. J. Bact. 34: 285-300 [167] 

545a. Ratcliffe, F. N., Myers, K., Fennessy, B. V., and Calaby, J. H. (1952) 
Myxomatosis in Australia. Nature 170: 7-11 [320] 

546. Reed, L. J., and Muench, H. (1938) A simple method of estimating fifty 

per cent endpoints. Am. J. 'Hyg. 27: 493-497 [43, 44] 

547. Reed, W,, and Carroll, J. (1902) The etiology of yellow fever; a supple- 

mentary note. Am. Med. 3: 301-305 [7] 

548. Reed, W., Carroll, J., Agramonte, A., and Lazear, J. W. (1911) Yellow 

fever: a compilation of various publications. 17. S. 61st Congress, Docum. 
No. 822, Washington [310] 

549. Reeves, W. C. ( 1945 ) Observations on the natural history of western 

equine encephalomyelitis. Proc. XLIX Meeting U. S. Livestock Sanitary 
Assoc. 150-158 [312] 

550. Remlinger, P. ( 1939) Quoted by Doenr in: Doerr and Hallauer (Ref. 187) 


551. Rhoades, M. M. (1946) Plastid mutations. Cold Spring Harbor Symp. 

Quant. Biol 11: 202-207 [356] 

552. Richards, O. W. ( 1946) Biological phase microscopy. Cold Spring Harbor 

Symp. Quant. Biol 11: 208-214 [60] 

553. Rickard, E. R., and Francis, T., Jr. (1938) The demonstration of lesions 

and virus in the lungs of mice receiving large intraperitoneal inoculations 
of epidemic influenza virus. /. Exp. Med. 67: 953-972 [281] 

554. Rivers, T. M. (1936) Immunity in virus diseases with particular reference 

to poliomyelitis. Am. /. Publ. Health 26: 136-142 [132] 

555. Rivers, T. M. (1937) Viruses and Koch's postulates. /. Bact. 33:1-12 


556. Rivers, T. M. (1942) Immunity in virus infections. Science 95: 107-112 

[132, 280] 

557. Rivers, T. M., and Pearce, L. (1925) Growth and persistence of filterable 

viruses in a transplantable rabbit neoplasm. J. Exp. Med. 42: 523-537 

558. Rivers, T. M., and Tillett, W. S. (1924) Further observations on the 

phenomena enountered in attempting to transmit varicella to rabbits. 
/. Exp. Med. 39: 777-802 [36, 340] 

559. Rivers, T. M., Haagen, E., and Muckenfuss, R. S. (1929) A study of vac- 

cinal immunity in tissue cultures. /. Exp. Med. 50: 673-685 [281] 

560. Rose, S. M., and Rose, F. C. (1952) Tumor agent transformations in 

amphibia. Cancer Res. 12: 1-12 [325, 332, 333, 341] 

561. Ross, A. F., and Stanley, W. M. (1938) The partial reactivation of 

formalized tobacco mosaic virus protein. /. Gen. Physiol. 22: 165-191 

561a. Ross, A. F., Rochow, W. F., and Siegel, B. M. (1952) High concentra- 
tions of virus X in plants doubly infected with potato viruses X and Y. 
Phytopath. 42: 473 [279] 

562. Rous, P. (1911) Transmission of a malignant new growth by means of a 

cell-free filtrate. /. Am. Med. Assoc. 56: 198 [7, 324] 

563. Rous, P. ( 1943 ) The nearer causes of cancer. /. Am. Med. Assoc. 122: 

573-581 [323, 325] 

406 Bibliography 

564. Rous, P. (1946) Concerning the cancer problem. Am. Scientist 34: 329- 

358 [327, 331, 332, 342] 

565. Rous, P., and Kidd, J. G. (1938) The carcinogenic effect of a papilloma 

virus on the tarred skin of rabbits. 7. Exp. Med. 67: 399-428 [338] 

566. Rous, P., Kidd, J. G., and Smith, W. E. ( 1952) Experiments on the cause 

of the rabbit carcinomas derived from virus-induced papillomas. /. Exp. 
Med. 96: 159-174 [328, 337] 

567. Rous, P., McMaster, P. D., and Hudack, S. S. (1935) The fixation and 

protection of viruses by the cells of susceptible animals. 7. Exp. Med. 
61: 657-688 [228] 

568. Ruska, H., and Kausche, G. A. ( 1943) Uber Form, Grosserverteilung und 

Struktur einiger Virus-Elementarkorper. Zentr. Bakt., Parasit., /, Orig. 
150:311-318 [64] 

569. Russell, D. S. (1932) The occurrence and distribution of intranuclear 

"inclusion bodies" in gliomas. 7. Path. Bad. 35: 625-634 [334] 

570. Sabin, A. B. (1935) The mechanism of immunity to filterable viruses. 

Br. J. Exp. Path. 16: 70-84 [128] 

571. Sabin, A. B. (1941) The filterable microorganisms of the pleuropneu- 

monia group. Bact. Rev. 5: 1-66 [4] 

572. Sabin, A. B. (1952) Nature of inherited resistance to viruses affecting the 

nervous system. Proc. Nat. Acad. Set. 38: 540-546 [30, 305] 

573. Sabin, A. B., and Ward, R. (1941) The natural history of human polio- 

myelitis. I. /. Exp. Med. 73: 771-793 [308] 

574. Salaman, R. N., and LePelley, R. H. (1930) Para-crinkle: a potato dis- 

ease of the virus group. Proc. Roy. Soc. (London) B, 106: 140-175 

575. Salk, J. E. (1944) A simplified procedure for titrating hemagglutinating 

capacity of influenza-virus and the corresponding antibody. /. Immunol. 
49: 87-98 [250, 252] 

576. Sanarelli, G. (1898) Das myxomatogene Virus. Beitrag 7Aim Studium 

der Krankheitserreger ausserhalb des Sichtbaren. Zentr. Bakt., Parasit., 
/, Orig. 23: 865-873 [326] 

576a. Sanders, F. K. (1953) Multiplication cycles in neurotropic viruses. In: 
The Nature of Virus Multiplication, Cambridge University Press [269] 

577. Sanders, M. (1939) Cultivation of the viruses. A critical review. Arch. 

Path. 28: 541-586 [224] 

577a. Scherer, W. F., and Syverton, J. T. ( 1952) Studies on the propagation in 
vitro of poliomyelitis viruses. III. /. Exp. Med. 96: 389-400 [227] 

577b. Scherer, W. F., Syverton, J. T., and Gey, G. O. (1953) Studies on the 
propagation in vitro of poliomyelitis viruses. IV. Viral multiplication in a 
stable strain of human malignant epithelial cells (Strain Hela) derived 
from an epidermoid carcinoma of the cervix. /. Exp. Med. 97: 695-710 

578. Schlesinger, M. (1932) Ober die Bindung des Bakteriophagen an homo- 

loge Bakterien. I, II. Z. Hyg. Infectionskrh. 114: 130-160 [164] 

579. Schlesinger, R. W. (1948) Viral diseases of the central nervous system. 

Scope and limitations of specific diagnostic methods. Pediatrics 1: 625- 
633 [305] 

Bibliography 407 

580. Schlesinger, R. W. ( 1949) The mechanism of active cerebral immunity to 

equine encephalomyelitis virus. II. /. Exp. Med. 89; 507-527 [282] 

581. Schlesinger, R. W. (1949) Local immune response and viral growth rate 

as factors in immunity to equine encephalomyelitis virus. Fed Proc 
8: 622-624 [131, 282] 

582. Schlesinger, R. W. (1950) Incomplete growth cycle of influenza virus in 

mouse brain. Proc. Soc. Exp. Biol. Med. 74: 541-548 [31 223 246 

583. Schlesinger, R. W. ( 1950) Interference between animal-pathogenic viruses. 

In: Viruses 1950, Pasadena, Cal. Inst. Technol. Bookstore [280, 283] 

584. Schlesinger, R. W. (1951) Studies on interference between influenza and 

equine encephalomyelitis viruses. Arch. ges. Virusforsch. 4: 501-517 

585. Schlesinger, R. W., Olitski, P. K., and Morgan, I. M. (1944) Induced re- 

sistance of the central nervous system to experimental infection with 
equine encephalomyelitis virus. III. /. Exp. Med. 80: 197-211 [283] 

586. Schnitzer, R. J., Buck, M., and Steiger, N. (1951) Chemotherapeutic 

effect of 2-hydroxy-l,4-naphthoquinonimine on infections of mice with 
Col SK virus. Proc. Soc. Exp. Biol. Med. 77: 182-187 [223] 

587. Schramm, G. (1947) Ober die Spaltung des Tabakmosaik- Virus und die 

Wiedervereiningung der Spaltstucke zu hohermolekularen Proteines. I. 
Die Spaltungreaktion. Z. Naturforsch. 2b: 112-121 [96, 101, 142] 

588. Schramm, G., and Miiller, H. (1940) Zur Chemie des Tabakmosaikvirus. 

Ober die Einwirkung von Keten und Phenylisocyanat auf das Virusprotein. 
Z. physiol Chem. 266: 43-55 [144] 

589. Seiffert, G. (1937) Ober das Vorkommen filtrabler Mikro-organismen in 

der Natur und ihre Ziichtbarkeit. Zentr. Bakt., Parasit., 1, Orig. 139: 337- 
342 [4] 

590. Selbie, F. R., Robinson, R. H. M., and Shope, R. E. (1948) Shope papil- 

loma virus: reversion of adaptation to domestic rabbit by passage through 
cottontail Br. J. Cancer 2: 375-380 [327] 

591. Sharp, D. ., Eckert, E. A., Beard, D., and Beard, J. W. (1952) Morphol- 

ogy of the virus of avian erythromyeloblastic leucosis and a comparison 
with the agent of Newcastle disease. /. Bad. 63: 151-161 [82] 

592. Sharp, D. G., Hook, A. E., Taylor, A. R., Beard, D., and Beard, J. W. 

(1946) Sedimentation characters and pH stability of the T2 bacterio- 
phage of Escherichia coli. J. Biol. Chem. 165: 259-270 [79] 

593. Sharp, D. J., Taylor, A. R., McLean, I. W., Jr., Beard, D., and Beard, J. W. 

( 1944 ) Density and size of influenza virus A ( PR8 strain ) in solution. 
Science 100: 151-153 [76] 

594. Sharpless, G. R., Da vies, M. C., and Cox, H. R. ( 1950) Antagonistic action 

of certain neurotropic viruses toward a lymphoid tumor in chickens with 
resulting immunity. Proc. Soc. Exp. Biol. Med. 73: 270-275 [306] 

595. Sheffield, F. M. L. (1939) Micrurgical studies on virus-infected plants. 

Proc. Roy. Soc. (London) B, 126: 529-538 [214] 

596. Shepard, C. G., and Wyckoff, R. W. G. ( 1946) The nature of the soluble 

antigen from typhus rickettsiae. U. S. Publ Health Rep. 61:761-767 

408 Bibliography 

597. Shope, R. E. (1932) A filtrable virus causing a tumor-like condition in 

rafebits and its relationship to virus myxomatosum. /. Exp. Med. 56: 803- 
822 [325] 

598. Shope, R. E. (1933) Infections papillomatosis of rabbits. /. Exp. Med. 

58: 607-624 [326, 335] 

599. Shope, R. E. (1935) Experiments on the epidemiology of pseudorabies. 

I. J. Exp. Med. 62: 85-99 [309] 

600. Shope, R. E. (1937) Immunization of rabbits to infectious papillomatosis. 

/. Exp. Med. 65: 219-231 [327] 

601. Shope, R. E. (1943) Swine influenza. In: Virus Diseases, Ithaca, Cornell 

University Press [7, 314] 

602. Shope, R. E. (1950) "Masking," transformation, and interepidemic sur- 

vival of animal viruses. In: Viruses 1950, Pasadena, Cal. Inst. Technol. 
Bookstore [30, 299, 327, 335, 361] 

602a. Shope, R. E. ( 1953) An antiviral substance from Penicillium funiculosum. 
I, II, III. 7. Exp. Med. 97: 601-650 [223] 

603. Sigel, M. M., Girardi, A. J., and Allen, E. G. (1951) Studies on the 

psittacosis-lymphogranuloma group. I. J. Exp. Med. 94: 401-413 

603a. Siminovitch, L., and Rapkine, S. (1951) Modifications biochimiques au 
cours du developpement des bacteriophages chez une bacterie lysogene. 
C. r. acad. sci. 232: 1603-1605 [182] 

604. Sinnott, E. W., Dunn, L. C,, and Dobzhansky, T. (1950) Principles of 

Genetics, 4th Ed., New York, McGraw-Hill [350] 

605. Smadel, J. E. (1952) Serologic reactions in viral and rickettsial infections. 

In: Rivers, Viral and Rickettsial Infections of Man, 2nd Ed., Philadelphia, 
Lippincott-- [133] 

606. Smadel, J. E., and Hoagland, C. L. ( 1942) Elementary bodies of vaccinia. 

Bact. Rev. 6: 79-110 [53, 54, 110, 120, 130, 348] 

607. Smith, E. F. (1916) Studies on the crowngall of plants: Its relation to 

human cancer. J. Cancer Res. 1: 231-310 [339] 

608. Smith, J. D., and Stocker, M. G. P. (1951) The nucleic acids of 

Rickettsia burneti. Br. J. Exp. Path. 32: 433-441 [366] 

609. Smith, K. M. ( 1941 ) Some notes on the relationship of plant viruses with 

vector and non-vector insects. Parasitology 33: 110-116 [317] 

610. Smith, K. M. (1950) An Introduction to the Study of Viruses, New York, 

Pitman [1] 

611. Smith, K. M., and Wyckoff, R. W. G. (1951) Electron microscopy of in- 

sect viruses. Research 4: 148-155 [83, 110] 

612. Smith, M. G., Blattner, R. J., and Keys, F. M. (1946) St. Louis encepha- 

litis. /. Exp. Med. 84: 1-6 [312] 

612a. Smith, M. H. D. (1952) The Berry-Dedrick transformation of fibroma 
into myxoma in the rabbit. Ann. N. Y. Acad. Sci. 54: 1141-1152 

613. Smith, W., Westwood, M. A., Westwood, J. C. N., and Belyavin, G. ( 1951 ) 

Spontaneous mutation of influenza virus A during routine' egg passage. 
Br. J. Exp. Path. 32: 422-432 [301] 

614. Snyder, J. C., Yeomans, A., Clements, D. H., Murray, E. S., Zarafonetis, C. J. 

D., and Tierney, A. ( 1947 ) Further observations on the treatment of 

Bibliography 409 

typhus fever with para-aminobenzoic acid. Ann. Int. Med. 27: 127 

615. Solomon, A. K. (1950) Cancer: Biophysics. In: Glasser, New Medical 

Physics., Chicago, Year Book Publishers [323] 

616. Sonneborn, T. M. (1949) Beyond the gene. Am. Scientist 37:33-59 


617. Sonneborn, T. M. (1950) The cytoplasm in heredity. Heredity 4: 11-36 

[355, 357] 

618. Soper, F. L. (1938) Situation de la fievre jaune au Bresil. Bull. Off. Int. 

Emat. Hyg. Publ 30: 1205-1208 [313] 

618a. Spencer, E. L., and Price, W. C. (1943) Accuracy of the local lesion 
method for measuring virus activity. I. Am. J. Bot. 30: 280-290 [47] 

619. Spiegelman, S. (1946) Nuclear and cytoplasmic factors controlling en- 

zymatic adaptation. Cold Spring Harbor Symp. Quant. Biol. 11: 256-277 

620. Spiegelman, S., and DeLorenzo, W. F. (1952) Substrate stabilization of 

enzyme-forming capacity during the segregation of a heterozygote. Proc. 
Nat. Acad. Sci. 38: 583-592 [358] 

621. Spiegelman, S., Lindegren, C. C., and Lindegren, G. (1945) Maintenance 

and increase of a genetic character by a substrate-cytoplasmic interaction 
in the absence of the specific gene. Proc. Nat. Acad. Sci. 31: 95-102 

621a. Stanley, W. M. (1934) Chemical studies on the virus of tobacco mosaic. 
I. Phytopath. 24: 1055-1085 [137] 

622. Stanley, W. M. (1935) Isolation of a crystalline protein possessing the 

properties of tobacco-mosaic virus. Science 81: 644-645 [92, 93] 

623. Stanley, W. M. (1937) Chemical studies on the virus of tobacco mosaic. 

/. Biol Chem. 121: 205-217 [209] 

624. Stanley, W. M. (1943) Chemical structure and the mutation of viruses. 

In: Virus Diseases, Ithaca, Cornell University Press [103, 137, 144] 

625. Steere, R. L. (1952) Virus increment curves obtained from counts of 

particles in clarified plant juice. Am. /. Bot. 39: 211-220 [211, 212] 
625a. Steere, R. L., and Williams, R. C. (1953) Identification of crystalline 
inclusion bodies extracted intact from plant cells infected with tobacco 
mosaic virus. Am, 7. Bot. 40: 81-84 [96] 

626. Steinhaus, E. A. (1946) Insect Microbiology, Ithaca, Comstock [347, 

364, 370] 

627. Steinhaus, E. A. (1949) Nomenclature and classification of insect viruses. 

Bact. Rev. 13: 203-223 [83] 

627a. Stent, G., and Maal0e, O. (1953) Radioactive phosphorus tracer studies 
on the reproduction of T4 bacteriophage. II. Biochem. Biophys. Acta 10: 
5,5-69 [176] 

628. Stent, G. S., and Wollman, E. L. ( 1950) Studies on activation of T4 bac- 

teriophage by cof actor. II. Biochem. Biophys. Acta 6: 307-316 [167] 

629. Stephenson, F. J., Schultz, E. S., and Clark, C. F. (1939) Inheritance of 

immunity from virus X (latent mosaic) in the potato. Phytopath. 29: 362- 
365 [24] 

630. Stone, J. D. (1947) Prevention of virus infection with enzyme of 

V. cholerae. I. Austral J. Exp. Biol Med. Sci. 26: 49-64 [257] 

410 Bibliography 

631. Stone, J. D., and Ada, G. L. (1950) Electrophoretic studies of virus-red 

cell interaction. Br. J. Exp. Path. 31: 275-284 [254] 

632. Storey, H. H. (1933) Investigations of the mechanism of the transmission 

of plant viruses by insect vectors. I. Proc. Roy. Soc. (London) B, 113: 
463-485 [33, 304, 315, 318] 

633. Strauss, M. J., Shaw, E. W., Bunting, H., and Melnick, J. L. (1949) 

"Crystalline" virus-like particles from skin papillomas characterized by 
intranuclear inclusion bodies. Proc. Soc. Exp. Biol. Med. 72: 46-50 
[105, 237] 

633a. Stuart-Harris, C. H. ( 1939 ) A neurotropic strain of human influenza 
virus. Lancet (1) 236:497-499 [266] 

634. Sugg, J. Y. ( 1949 ) An influenza virus pneumonia of mice that is non- 

transferable by serial passage. J. Bact. 57: 399-403 [271] 

635. Sugg, J. Y., and Magill, T. P. (1948) The serial passage of mixtures of 

different strains of influenza virus in embryonated eggs and in mice. /. 
Bact. 56: 201-206 [284, 288] 

636. Svedmyr, A. ( 1948 ) Studies on a factor in normal allantoic fluid inhibiting 

influenza virus haemagglutination. Occurrence, physicochemical prop- 
erties, and mode of action. Br. J. Exp. Path. 29: 295-308 [257, 258] 

637. Sy verton, J. T., and Berry, G. P. ( 1936 ) Coexistent infections of individual 

cells by more than one filterable virus. J. Bact. 32: 356-357 [238] 

638. Syverton, J. T., and Berry, G. P. (1941) Hereditary transmission of the 

Western type of equine encephalomyelitis virus in the wood tick, Derma- 
centor andersoni Stiles. /. Exp. Med. 73: 507-530 [311] 

639. Syverton, J. T., and Berry, G. B. (1947) Multiple virus infection of single 

host cells. J. Exp. Med. 86: 145-152 [284] 

640. Takahashi, W. - N. (1941) Changes in nitrogen and virus content of de- 

tached tobacco leaves in darkness. Phijtopath. 31: 1117-1122 [210] 
641.. Takahashi, W. N., and Rawlins, T. E. (1933) Rod-shaped particles in 
tobacco mosaic virus demonstrated by stream double refraction. Science 
77: 26-27 [94] 

641a. Tall, M. G., Price, W. C., and Wertman, K. (1949) Differentiation of 
tobacco and tomato ringspot viruses by cross immunization and comple- 
ment fixation. Phytopath. 39: 288-299 [127] 

642. Tamm, L, and Horsfall, F. L., Jr. (1950) Characterization and separation 

of an inhibitor of viral hemagglutination present in urine. Proc. Soc. Exp. 
Biol Med. 74: 108-114 [257] 

643. Tang, F. F., and Wei, H. (1937) Morphological studies on vaccinia virus 

cultivated in the developing egg. J. Path. Bact. 45: 317-323 [233] 

644. Taylor, A. (1943) The successful production of a mammalian tumor with 

a virus-like principle. Science 97: 123 [330] 

645. Taylor, A. R. (1946) Chemical analysis of the T2 bacteriophage and its 

host, Escherichia coli (strain B). /. Biol Chem. 165:271-284 [103] 

646. Theiler, M. (1937) Spontaneous encephalomyelitis of mice, a new virus 

disease. J. Exp. Med. 65: 705-719 [37] 

647. Theiler, M., and Smith, H. H. (1937) The effect of prolonged cultiva- 

tion in vitro upon the pathogenicity of yellow fever virus. J. Exp. Med. 
65: 767-786 [300] 

Bibliography 411 

648. Thompson, R. L., Price, M. L,, Minton, S. A., Jr., Elion, G. B., and Hitchings, 

G. H. ( 1950) Effects of purine derivatives and analogues on multiplica- 
tion of the vaccinia virus. /. Immunol 65: 529-534 [230] 

648a. Thompson, R. L., Price, M., Minton, S. A., Jr., Falco, E. A., and Hitchings, 
G. H. (1951) Protection of mice against the vaccinia virus by the ad- 
ministration of phenoxythiouracils. J. Immunol. 67 ': 483-491 [223] 

648b. Thung, T. H. ( 1931) Quoted by Price (Ref. 533) [276] 

649. Trager, W. (1935) Cultivation of the virus of grasserie in silkworm tissue 

cultures. /. Exp. Med. 61: 501-513 [237] 

650. Trager, W. ( 1938) Multiplication of the virus of equine encephalomyelitis 

in surviving mosquito tissues. Am. J. Trop. Med. 18: 387-393 [312] 

651. Traub, E. (1939) Epidemiology of lymphocytic chorfomeningitis in a 

mouse stock observed for four years. J. Exp. Med. 69: 801-817 [38, 
309, 340, 361] 

652. Twombly, G. H., and Meisel, D. ( 1946) The growth of mammalian tumors 

in fertile eggs. Is a filterable cancer virus produced? Cancer Res. 6: 82- 
91 [330] 

653. Twort, F. W. (1915) An investigation on the nature of ultramicroscopic 

viruses. Lancet (2), 189: 1241-1243 [7] 

654. Uppal, B. M. (1934) The movement of tobacco mosaic virus in leaves of 

Nicotiana sylvestris. Indian J. Agr. Sci. 4: 865-873 [216] 
654a. Vago, C. ( 1951 ) Phenomenes de "latentia" dans une maladie a ultra virus 

des insectes. Rev. can. biol 10: 299-308 [320, 349] 
654b. Valleau, W. D. (1946) Breeding tobacco varieties resistant to mosaic. 

Phytopath. 36: 412 [304] 
654c. Valleau, W. D. (1952) The evolution of susceptibility to tobacco mosaic 

in Nicotiana and the origin of tobacco mosaic virus. Phytopath. 42: 40-42 


655. Van Rooyen, C. E., and Rhodes, A. J. ( 1948) Virus Diseases of Man, 2nd 

Ed., New York, Nelson [29, 58, 224, 241] 

656. Vilches, A., and Hirst, G. K. (1947) Interference between neurotropic 

and other unrelated viruses. /. Immunol. 57: 125-140 [266, 284] 

657. Visconti, N., and Delbriick, M. (1953) The mechanism of genetic recom- 

bination in phage. Genetics 38: 5-33 [190, 193] 

658. Volkerts, M., and Horsfall, F. L., Jr. (1947) Studies on a lung tissue 

component which combines with pneumonia virus of mice ( PVM ) . J. Exp. 
Med. 86: 393-407 [260] 

659. Von Magnus, P. ( 1951 ) Propagation of the PR8 strain of influenza A 

virus in chick embryo. I. Acta Path. Microb. Scand. 28: 250-277 [261, 

660. Von Magnus, P. ( 1951 ) Propagation of the PR8 strain of influenza A virus 

in chick embryo. II. Acta Path. Microb. Scand. 28: 278-293 [265] 

661. Wahl, R., and Blum-Emerique, L. (1949) Purification et concentration du 

bact&iophage. II. Ann. Inst. Pasteur 76: 103-121 [103] 

662. Wallace, J. M. (1944) Acquired immunity from curly top in tobacco and 

tomato. J. Agr. Res. 69: 187-214 [7, 279] 

663. Wallace, J. M. (1950) Immunological properties of plant viruses. In: 

Viruses 1950, Pasadena, Cal. Inst. Technol. Bookstore [279] 

664. Warren, J., Weil, M. L., Russ, S. B., and Jeffries, H. ( 1949) Purification of 

412 Bibliography 

certain viruses by use of protamine sulfate. Proc. Soc. Exp. Biol. Med. 
72:662-664 [88, 105] 

665. Watson, J. D. (1950) The properties of X-ray inactivated bacteriophage. 

I. /. Bact. 60:697-718 [135, 149, 168, 194] 

666. Watson, J. D. ( 1952 ) The properties of X-ray inactivated bacteriophage. 

II. /. Bact. 63: 473-485 [149] 

666a. Watson, J. D., and Crick, F. H. C. (1953) Genetical implications of the 
structure of deoxyribonucleic acid. Nature 171: 964-996 [100] 

667. Watson, M. A., and Roberts, F. M. (1940) Evidence against the hypothe- 

sis that certain plant viruses are transmitted mechanically by aphides. Ann. 
Applied Biol. 27: 227-233 [315] 

667a. Weed, L. L., and Cohen, S. S. ( 1951 ) The utilization of host pyrimidincs 
in the synthesis of bacterial viruses. J. Biol. Chem. 192: 693-700 [176] 

668. Weidel, W. (1951) Ober die Zellmembran von Escherichia coli B. Z. 

Naturforsch. 6b: 251-259 [168] 

669. Weigle, J. J., and Delbriick, M. (1951) Mutual exclusion between an in- 

fecting phage and a carried phage. /. Bact. 62: 301-318 [189, 202, 

670. Weil, M. L., Beard, D., Sharp, D. C., and Beard, J. W. (1948) Purifica- 

tion, pH stability and culture of the mumps virus. J. Immunol. 60: 561- 
582 [245, 248] 

671. Weiss, E. (1949) The extracellular development of agents of the psittaco- 

sis-lymphogranuloma group ( Chlamydozoaceae ) . J. Infectious Diseases 
84: 125-149 [232] 

672. Weller, T. H., Robbins, F. C., and Enders, J. F. (1949) Cultivation of 

poliomyelitis virus in cultures of human foreskin and embryonic tissues. 
Proc. Soc. Exp. Biol. Med. 72: 153-155 [227] 

673. Wertman, K. (1948) The Weil-Felix reaction. In: The Rickettsial Dis- 

eases of Man, A.A.A.S., Washington [368] 

674. Weyer, F., Friedrich-Freksa, H., and Bergold, G. ( 1944) Die Beziehungen 

der Rickettsien zu Bakterien und Viren. Naturwiss. 32: 361-365 [366] 

675. Wildman, S. G., Cheo, C. C., and Bonner, J. (1949) The proteins of 

green leaves. III. J. Biol. Chem. 180: 985-1001 [209, 210] 

676. Wilkins, M. H. F.,, A. R., Seeds, W. E., and Oster, G. (1950) 

Tobacco mosaic virus crystals and tri-dimensional microscopic vision. 
Nature 166: 127-139 [96] 

676a. Williams, R. C. (1952) High-resolution electron microscopy of the 
particles of tobacco mosaic virus. Biochem. Biophys. Acta 8: 227-244 

677. Williams, R. C., and Backus, R. C. (1949) Macromolecular weights de- 

termined by direct particle counting. I. J. Am. Chem. Soc. 71: 4052- 
4057 [67] 

678. Williams, R. C., and Steere, R. L. ( 1951 ) Electron microscopic observa- 

tions on the unit of length of the particles of tobacco mosaic virus. /. Am. 
Chem. Soc. 73: 2057-2061 [94] 

679. Williams, R. C., and Wyckoff, R. W. G. (1946) Applications of metallic 

shadow-casting to microscopy. /. Appl. Phys. 17: 23-33 [61] 

680. Williams, R. C., Backus, R. C., and Steere, R. L. ( 1951 ) Macromolecular 

Bibliography 413 

weights determined by direct particle counting. II. /. Am. Chem. Soc. 
72: 2062-2066 [68] 

681. Whitman, L. (1937) The multiplication of the virus of yellow fever in 

Aedes aegypti. J. Exp. Med. 66: 133-143 [311] 

682. Wolbach, S. B. (1948) The pathology of the rickettsial diseases of man. 

In: The Rickettsial Diseases of Man, A.A.A.S., Washington [369] 

683. Wollman, E. (1927) Recherches sur la bacteriophagie (Phenomene de 

Twort-d'Herelle). Ann. Inst. Pasteur 41: 883-918 [350] 

684. Wollman, E., and Lacassagne, A. (1940) Evaluation des dimensions des 

bacteriophages au moyen des rayons X. Ann. Inst. Pasteur 64: 5-39 

685. Woodruff, A. M., and Goodpasture, E. W. ( 1931) Susceptibility of chorio- 

allantoic membrane of chick embryo to infection with fowl-pox virus. 
Am. 7. Path. 7: 209-222 [239] 

686. Woods, M. W., and DuBuy, H. G. (1943) Evidence for the evolution of 

phytopathogenctic viruses from mitochondria and their derivatives. I. 
Phytopath. 33: 637-655 [213] 

687. Wright, S. ( 1941) The physiology of the gene. Physiol Rev. 21: 487-527 

687a. Wyatt, G. R. (1952) Specificity in the composition of nucleic acids. 

Expt. Cell Res., Suppl. 2: 201-215 [97] 
687b. Wyatt, G. R., and Cohen, S. S. ( 1952) A new pyrimidine-base from bac- 

teriophage nucleic acids. Nature 170: 1072-1073 [105, 174, 180] 

688. Wyckoff, R. W. G. (1949) Electron Microscopy, New York, Interscience 

[60, 91, 92, 93] 

689. Wyckoff, R. W. G. (1951) The virus of \accinia in chick embryo mem- 

brane. Proc. Nat. Acad. Set. 37: 565-569 [234] 

690. Wynd, F. L. ( 1943 ) Metabolic phenomena associated with virus infection 

in plants. But. Rev. 9: 39.5-465 [218] 
690a. Yamafuji, K. (1952) Mechanism of artificial virus formation in silkworm 

tissues. Enzymologia 15: 223-231 [320, 349] 
690b. Yarwood, C. E. (1952) Latent period and generation time for two plant 

viruses. Ann. /. Bot. 39: 613-618 [212] 

691. Youden, W. J. (1937) Use of incomplete block replications in estimating 

tobacco-mosaic virus. Contrib. Boyce Thompson Inst. 9: 41-48 [47] 

692. Youden, W. J., and Bealc, H. P. (1934) A statistical study of the local 

lesion method for estimating tobacco mosaic virus. Contrib. Boyce Thomp- 
son Inst. 6: 437-454 [47] 

693. Youden, W. J., Beale, H. P., and Guthrie, J. D. ( 1935) Relation of virus 

concentration to the numbers of lesions produced. Contrib. Boyce Thomp- 
son Inst. 7: 37-53 [54] 

694. Zarafonetis, C. J. D. (1948) The serological reactions in the rickettsial 

diseases of man. In: The Rickettsial Diseases of Man, A.A.A.S., Wash- 
ington [366] 

695. Zinder, N. D., and Lederberg, J. (1952) Genetic exchange in Salmonella. 

J. Bact. 64: 679-699 [115, 206, 352] 

696. Zinsser, H. ( 1935) Rats, Lice, and History, Boston, Little, Brown [368] 

697. Zinsser, H., and Schoenbach, E. B. (1937) Studies on the physiological 

conditions prevailing in tissue cultures. /. Exp. Med. 66: 207-227 [229, 


Viruses are listed under the name of the corresponding disease, except where 
omitting the word "virus" would create confusion. 

Abutilon variegation, 14, 34 
Acquired immunity, 229 

plant viruses, 275 
Acriflavine, 357 
Activation energy, 139, 141 
Adaptation, 36, 290/ 
Adaptive enzymes, 179, 358 
Adenine, 97 
Aedes aegypti, 310/ 
Agallia constricta, 318 
Age and virus susceptibility, 30, 305 
Agglutination, 118 

virus-coated particles, 119 
Alastrim, 11, 298 
Alfalfa mosaic, 138 
Allantoic fluid, normal components, 

108, 248 
Allantoic membrane, virus synthesis in, 

Allantois, 240 

inoculation, 241 
Allele, 190 
Alpha particles, 147 
Amino acids, analogues, 146 
bacteriophage, 104 
inhibition by, 230 
tobacco mosaic, 102 
Amnion, 240 

inoculation, 242 
Anaphylactic reactions, 88 
Animal viruses, composition, 105 
growth, 22 If 
inclusions, 230/ 
interference, 280/ 
maturation, 235 
plaques, 42 
serology, 128/ 

transmission, arthropods, 35, 310 
mechanical, 309 

Animal viruses, variation, 298 

vectors, 310 
Antibiotics, 145, 223 

against plant viruses, 219 
Antibodies, definition, 116 
local, 282 
persistence, 30, 132 
plant, 279 
Antigen-antibody reactions, 117 
Antigens, bacterial, 167 
bacteriophage, 120 
complement-fixing, influenza, 263 
definition, 116 
host, 107, 260 
Paramecium aurelia, 357 
vaccinia, 120, 130 
See also Soluble antigens 
Aphthous fever, see Foot-and-mouth 
Aphthous stomatitis, 38 
Arrhenius plot, 140 
Arthropod vectors, 7, 310/ 

rickettsiae, 364, 369 
Aster yellows, 14, 27, 36 
heat inactivation, 138 
multiplication in vector, 318 
transmission, 317 
Asymmetry coefficient, 76 
Aucuba mosaic, 121, 277 
Aureomycin, 145, 223, 369 
Autosterilization, 280 
Autotrophs, 346 
Auxotrophs, 346 

Bacillus megatherium, 206 
Bacteria, genetic transfer, 351 
L forms, 4 

lysogenic, see Lysogenic bacteria 
phage-resistant mutants, 187, 303 
Bacterial enzymes, 178 




Bacterial virus, see Bacteriophagc 
Bacteriophage, adsorption, 164 

cations, 166 

cofactors, 166, 293 
amino acids, 104 
antigens, 120 
burst size, 168 
calcium requirement, 168 
classification, 18 
complement fixation, 126 
composition, 103 
depressor effect, 188 
desoxyribonucleic acid, 97, 105, 173, 


"doughnuts," 171 
eclipse period, 170 
ecology, 208 

genetic recombination, 189, 294, 350 
genetic transfer, 206, 353 
genetics, 183, 291 
ghost, 84, 104, 143, 173, 199 
heterozygotes, 193 
host nucleus, 180 
host syntheses, 177 
inactivation, bacterial extracts, 168 

formalin, 144 % 

heat, 138, 142 

radiation, 149, 151, 153, 165, 180, 

radioactive decay, 154 
infectious unit, 105 
interference, 188, 273f 
intracellular irradiation, 156, 195 
isotope transfer, 176 
latent period, 159, 168, 199 
linkage system, 191 
lysis, 199 

from without, 168, 179 

inhibition, 184, 292 

premature, 170 
lysogenic cycle, 158 
lytic cycle, 159 
maturation, 172 
mixed infection, 186, 274 
modification, 197 

host-induced, 197, 207, 295, 328 
morphology, 65/ 
multiple infection, 186 
multiplicity of infection, 164 

Bacteriophage, multiplicity reactivation, 
156, 195 

mutants, 183, 203, 291, 292, 293 

mutation, 183 

mutual exclusion, 188, 202, 273 

neutralization, 123 
reversibility, 125 

one-step growth experiment, 159, 161 

origin of constituents, 175 

osmotic shock, 173 

phenotypic mixing, 295 

photoreactivation, 155, 194 

plaques, 21, 41, 184, 188, 200 

proflavine effect, 126, 170, 199 

prophage, 200/ 

chromosomal location, 352 
in spores, 201 
loss, 205 

radiation, 149, 151, 153, 156, 165, 
180, 194, 195 

receptors, 167 

replication, 194 

resistance, 9, 186, 187, 303 

Salmonella, 167 

sedimentation, 79 

serology, 122/ 

single-burst experiment, 162, 192 

size, 57, 160 

skin (ghost), 84, 104, 143, 173, 199 

T group, 159, 160 

temperate, 159, 274 

titration, 41, 52 

typing, 167, 206 

ultrafiltrate factor, 111, 126, 131, 173 

variation, 291 

vegetative stage, 175, 185, 197 

virulent, 159, 273 

x-ray effects, 135 
Bacteroids, 347, 370 
Bartonella bacilliformis, 373 
Bean mosaic, 14, 217 
3,4-Benzopyrene, 323 
Beta rays, 147 
Biochemical evolution, 346 
Biotin, 108, 348 

Blue disease of Japanese beetle, 372 
Blue tongue of sheep, 223 
Bellinger bodies, 231 
Bombyx mori (silkworm), 81 



Borreliota, 18 
Brill's disease, 372 
Brown-Pearce carcinoma, 337 
Burst size, bacteriophage, 168 

Cacoecia murinana, 83 

Canary pox, 331 

Cancer, 32 If 

Capsular diseases of insects, see Granu- 


Carcinoma, 322/ 
Carcinogenic agents, 323 
Carp pox, 9 

Cation exchange resins, 89, 166 
Cattle plague, 12 
Cell components, relation to viruses, 

272, 348, 359 
Cell origin, 359/ 
Centrifugation, 73/ 

density gradient, 89 
Centrioles, 354 
Centrosome, 354 

Chemical agents, effects on viruses, 143 
Chemical composition of viruses, 96/ 
Chemotherapy, 144, 223 
Chick embryo, culture, 239/ 
development, 239 
vaccines, 248 

Chicken pox (varicella), 11, 24, 37 
inclusions, 237 
size, 57 

Chlamydozoa, 231 
Chlamydozoaceae, 18 
Chloromycetin, 145, 223, 369 
Chloroplasts, 213 
Cholesterol, 107 

Chorioallantoic membrane, develop- 
ment, 240 
grafts, 246 
inoculation, 242 

lesions (pocks), 25, 42, 129, 249 
titration, 42, 249 
Chorion (serosa),240 
Choristonenra fumiferana, 80, 82 
Chromosomes, 192, 272 
Cicadulina mbila, 304, 315 
Circulifer tenellus, 33, 279, 315 
Classification, 16, 278 
Clone, 185 

Clover club-leaf, 317 

Cocoa swollen shoot, 14 

Colicines, 207, 353 

Colorado fever, 311 

Complement fixation, 119, 126, 134 

See also names of viruses 
Condyloma acuminatum, 325 
Contagium virum fluidum, 6 
Copper, 108 

Corn streak, 14, 304, 316 
Corn stripe, 14 

Corynebacterium diphtheriae, 206 
Cowpox, see Vaccinia 
Coxiella burneti, 365/ 

popiHta, 372 

Coxsackie group, 10, 57 
Crossing-over, 192 

Cross-protection, 16, 18, 275, 276, 278 
Crown gall, 216, 339 
Crystallization of viruses, 90 
Cucumber mosaic, 14, 215, 297 

composition, 101 

serology, 121 
Culex tarsalis, 312 
Cyanide, 170, 229, 369 
Cytochrome oxidase, 258 
Cytoplasmic inheritance, 354/ 
Cytosine, 97, 105 

Dark-field microscope, 58 
DDT, 371 

Definition of viruses, 2 
Dengue, 10, 245, 285, 305, 310, 312 
Density, measurement, 76 
Density gradient centrifugation, 89 
Depressor effect, bacteriophage, 188 
Dermacentor andersonii, 310 
Dermacentroxenus akari, 365 
Dermacentroxenus pediculi, 365 
Dermacentroxenus rickettsi, 365f 
Dermanyssus gattinae, 312 
Desoxyribonuclease, 97, 143, 173, 179 
Desoxyribomicleic acid (DNA), 97, 

bacteriophage, 97, 105, 173, 174 

reduplication, 100 

structure, 100 

transforming principles, 351 
Deuterons, 147 



1,2,5,6-Dibenzanthracene, 323 
Diffusion, 74 

measurement, 79 

nonspherical particles, 75 
Diffusion constant, 75 
Dinitrophenol, 229 
Diplococcus pneumoniae, 351 
Diploid number, 193 
Distemper, 12, 288 
DNA, see Desoxyribonucleic acid 
Dodder, 34, 308, 309 
"Doughnuts," bacteriophage, 171 
Drosophila, CO2 sensitivity, 357 

sigma factor, 359 
Drying, 142 
Dual infection, animal viruses, 238, 284 

plant viruses, 215, 279 

Earthworm, 314 

Eclipse period, 170, 212, 222, 263, 269 

Ecology, of bacteriophage, 208 

of viruses, 312 
Ectoderm, 239 
Ectromelia, 301 

hemagglutination, 250 

morphology, 64 

titration, 249 

Egg transmission, 312, 361 
Egg-white inhibitor, 258 
Electron microscope, particle count, 67 

techniques, 60 
Electrophoresis, 79, 209 
Elementary bodies, definition, 6, 58 
Elm phloem necrosis, 14 
Enation mosaic, 121 
Enations, 28 

Encephalitis, Russian type, 306 
Encephalomyocarditis, 251 

Columbia MM, 285 

Columbia SK, 223, 285, 287 
Entoderm, 239 
Enzyme activity, MNI viruses, 253/, 

Enzymes, action on viruses, 143 

adaptive, 179, 358 

bacterial, 178 * 

in virus particles, 109 

receptor-destroying, 110, 256/ 

Equine encephalomyelitis, 10, 154, 230, 
234, 236 

antibodies, 282 

composition, 106 

egg culture, 244 

epidemiology, 312 

interference, 283, 285, 286, 287 

plaques, 26 

sedimentation, 78 

sensitivity, 305 

size, 57 

spread, 308 

tissue culture, 229 

vectors, 310 

multiplication in, 312 
Escherichia colt, 167, 202 

fertility agent, 352 

strain B, 22, 155, 295 
mutants, 303 

strain K-12, 205, 303 
DL-Ethionine, 230 
Evolution, regressive, 4 

rickettsiae, 370 

sexual mechanisms, 361 

virus diseases, 301 

viruses, 318 
Excitation by radiation, 147 

False air space, chick embryo cultures, 


Feline infectious enteritis, 12 
Feline pneumonitis, 62, 286 
Feulgen reaction, 97, 233 
Fibroma, 13, 14, 57, 228, 285, 325 

tarring, 338 
Fibroma-myxoma group, 299 

transformation, 302 
Fiji disease of sugar cane, 14, 216 
Filtration end point, 72 
Flow birefringence, 94 
Fluorescence microscope, 60 
Folic acid, 369 

Foot-and-mouth (aphthous fever), 6, 
12, 324, 331 

growth, 228 

hemagglutination, 251 

interference, 285 

inactivation by x-rays, 151 



Foot-and-mouth, size, 57 

tissue culture, 227 
Fowl leukemia, 9, 13, 108, 324, 332, 


Fowl plague, 12, 227, 234, 251 
Fowl pox, 13, 18, 246, 286, 331 
Fowl sarcomas, 79 

See also Fujinami sarcoma, Rous sar- 

Fowl tumors, 7, 15, 324/ 
Francis inhibitor, 258 
Frei test, 134 
Friction coefficient, 73 
Frog, kidney carcinoma, 9, 325, 333 
Fujinami sarcoma, 332, 336 

Gamma globulin, 132 

Gamma rays, 147 

Generation time, influenza, 221 

tobacco mosaic, 212 
Generalized reactions, 26/ 
Genes, 115 

naked, 350 

relation to viruses, 349 
Genetic parasitism, 353 
Genetic recombination, bacteriophage, 
189, 294, 350 

influenza, 267, 350 
Genetic systems, merging, 361 
Genetic transfer, 206, 351, 353 
German measles, 11, 246 
Glycerin, 144 
"Gradocol" membranes, 71 
Graft, transmission, 33, 308 

tumors, 322 

virus transmission by, 33, 115, 308 
Granuloses (capsular diseases of in- 
sects), 13, 57, 82, 83, 272 
Growth curve, animal viruses, 22 If 

bacteriophage, 159 

plant viruses, 210f 
Guarnieri bodies, 233, 234, 247, 298 
Guanine, 97 

Haploid number, 193 

Haptens, 116 

Heat inactivation, 138, 139, 141, 142 

Heat therapy, plant viruses, 145 

Helenine, 223 

Hemagglutination, 70, 250/ 

Hemagglutination, cation effects, 260 

elution, 253 

inhibitors, 257 
egg-white, 257 
Francis, 258 

receptor gradient, 254 
Hemolysis by viruses, 254 
Hemophilus influenzae, 351 
Hemophilus infiuenzae suis, 314 
Herpes (Herpes simplex), 10, 35, 37, 
238, 343 

epidemiology, 320 

immunity, 280 

inclusions, 237 

interference, 285, 286 

latency, 271 

Magrassi phenomenon, 282 

neurotropism, 300 

titration, 249 
Herpes zoster, 10, 57 
Heterogenesis, 343 
Heterotrophs, 346 
History of virology, 5 
Hit theory, 150 
Hog cholera, 12 
Host antigens, 107, 260 
Host-induced modifications, bacterio- 
phage, 197, 207, 295, 328 
Host specificity, 7, 227 
Hydration, 76 

5-Hydroxymethylcytosine, 97, 105, 174, 

Immunity, 131/, 280 

lysogenic bacteria, 30, 200/ 

tissue, 280 

Inclusion conjunctivitis, 10 
Inclusions, 6 

animal viruses, 230/, 247, 298 
intranuclear, 237 

plant viruses, 214, 215 

intranuclear, 215 
Incubation period, 220 

specificity, 29 

titration by, 45 

vectors, 311 
Indole, 166 
Induction, 201, 328 
Infection, route of, 31 



Infectious epithelioses, 324 
Infectious hepatitis, 10 
Infectious unit, 40/ 

bacteriophage, 105 

relation to virus particle, 52, 54 

statistical interpretation, 48 
Influenza, 11, 15 

adsorption and elution in lungs, 254 

antibodies, 132 

antigen, soluble, 110, 130, 263/ 

composition, 106, 107 

generation time, 221 

genetic recombination, 267, 350 

growth, 221, 245, 261 

inhibition by polysaccharides, 146 

hemagglutination, 70, 250/ 
electron microscopy, 67 

heat inactivation, 141, 142 

immature forms, 263, 288 

immunity, 281 

in mouse brain, 265 

interference, 261, 283/ 

isolation, 248 

isotope labeling, 249 

morphology, 63, 64, 82, 267 

multiplicity reactivation, 268 

neurotropism, 24 % 6, 266, 267 

O-D variation, 253, 300 

one-step growth experiments, 261 

purification, 89 

radiation effects, 155 

serology, 129, 130 

size, 57 

swine, see Swine influenza 

tissue cultures, 227 

titration, 40, 45 

toxicity, 31, 223, 266, 271 

types, 299 

vaccines, 133 

Von Magnus phenomenon, 265, 288 

See also MNI viruses 
Insect transmission, 314 

artificial, 33 
Insect vectors, 310 

genetic factors, 316 

See also Arthropod vectors; Vectors 
Insect viruses, 9, 19, 272 

composition, 110 

polyhedra, 237 

Insect viruses, see also Granuloses; Poly- 
hedral diseases 
Interference, 261, 266, 273/ 

animal viruses, 280/ 
inactive, 284 

bacteriophages, 188, 273/ 

mechanism, 287 

plant viruses, 275/ 

tissue cultures, 283 
Intracellular irradiation, 156, 195 

See also Radiation 
Isotope tracers, 175, 210, 213, 249 

Japanese (type B) encephalitis, 10, 57 

Kappa factor, 356, 359 
Kinetosomes, 355 
King Edward potato, 359 
Klebsiella pneumoniae, 222, 259 
Koch's postulates, 32 

Laryngotracheitis, 9, 13, 249, 286 
Latency, 38, 340, 360 
Latent dodder mosaic, 37 
Latent viruses, 36, 271, 349 

and tumors, 340 
Layering phenomenon, 94 
Lettuce mosaic, 217 
Levels of integration, 113 
Liquid crystals, 94 
Light absorption, 147 
Light scattering, 79 
Little peach, 278 
Living organism, definition, 114 
Local reactions, 21 
Logarithms, 40 
Louping ill, 12, 57, 129, 238, 285, 286, 


LS antigen, 130 
Lungworm, 314 

Lymphocytic choriomeningitis, 10, 131, 
285, 286, 309, 343 

latency, 38, 340, 361 
Lymphogranuloma venereum, 11, 229, 
285, 286 

epidemiology, 309 

Frei test, 134 

growth, 247 
Lysines, 199 



Lysis, 199 

from without, 168, 179 

inhibition, 184, 292 

premature, 170 
Lysogenesis, 200/ 
Lysogenic bacteria, 30, 200/ 

immunity, 201, 202 

induction, 201, 328 

superinfection, 204 

toxicity, 206 
Lysogenic cycle, bacteriophage, 158 

Macrophages, 228 

Macrosteles divisus, 318 

Mad itch ( pseudorabies ) , 12, 30, 305, 

309, 320 

Magrassi phenomenon, 282 
Major host, 8 
Malonate, 229 
Mare abortion, 246 
Masking, 296, 335/ 

rabbit papilloma, 327, 361 

swine influenza, 314 
Maternal inheritance, 355 
Maturation, bacteriophage, 172 

viruses, 272 
Measles, 11, 24, 25, 132, 309 

epidemiology, 320 

incubation period, 29 
Melanophores, 359 
Meningopneumonitis, 269 
Mesoderm, 239 
Metastases, 321 
Methionine analogues, 230 
Methylcholanthrene, 323 
5-Methylcytosine, 97 
5-Methyltryptophan, 166 
Mh unit, 56 
Microsomes, 105, 354 
Milk factor, 57, 306, 328/, 340 
Mites, 312 

Mitochondria, 214, 237, 272, 354 
MNI viruses (Mumps, Newcastle, in- 
fluenza), 18, 251/, 284 
Molecular distances, 56 
Molecular weight, 56 
Molecule, 112 

Molluscum contagiosum, 11, 18, 231, 
232, 238 

Molluscum contagiosum, inclusions, 
234, 236 

morphology, 59 

size, 57 
Mosaic diseases, 218 

See also names of diseases 
Mouse breast carcinoma, 328 
Mouse encephalomyelitis (Theiler's), 
37, 45 

growth, 269 

hemagglutination, 251 

interference, 285 

morphology, 62 

size, 57 

Mouse leukemia, 330 
Mucin, 258 

Multiplicity reactivation, 156, 195, 268 
Mumps, 11, 29, 245, 247, 285, 286, 287 

hemagglutination, 250/ 

hemolysis, 254 

incubation period, 29 

reproduction, 269 

skin test, 134 

See also MNI viruses 
Murray Valley encephalitis, 25 
Mustard gas, 179 
Mutability, 7, 290/, 345 

plastids, 356 
Mutation rate, 291 
Mutations, 291/ 

and classification, 17 

bacteriophage, 183 

fluctuation test, 303 

host adaptation, 36 

in evolution, 319 

pleiotropic, 300 

reverse, 294 

somatic, 338 
Mutual exclusion, bacteriophage, 188, 

202, 273 
Mycetomes, 370 
Mycobacterium leprae, 3 
Myxoma, 13, 25, 249, 298, 325 

rabbit control by, 320 

serology, 129 

size, 57 

transformation, 302 
Myzus persicae, 315 
Myzus pseudosolani, 316 



Naked genes, 350 
Nature of viruses, 112, 344/ 
Necrotic lesions, 22, 23, 47 
genetic influences, 22, 304 
morphology, 21 
titration by, 46 
Negri bodies, 25, 230 
Neoplasms, 32 If 
Nerve transport, 308 
Newcastle disease, 9, 13, 244, 285, 286, 

287, 300 

hemagglutination, 250/ 
hemolysis, 254 
morphology, 82 
toxicity, 223 
See also MNI viruses 
Nicotiana, response to tobacco mosaic, 

22, 304, 319 
Nicotiana digluta, 304 
Nicotiana glutinosa, 23, 47, 304 
Nicotiana silvestris, 277 
Nicotiana tabacum, 23, 27, 304 

variety Ambalema, 304 
Nitrogen mustard, 182 
Nucleic acids, 97, 141 
ionizing radiations, 152 
ultraviolet absorption, 147 
Nucleoproteins, 99 

and virus nature, 345 
Numerical aperture, 58 
Nutritional deficiencies, 31, 223 

Obligate parasites, 3 

Oncolysis, 306 

One-step growth experiments, 159, 161, 


Oral papillomatosis, 326 
Origin of cell, 359/ 
Origin of viruses, 344/ 
Ornithosis, 9 
Orroya fever, 373 

Panagglutinability, 253 
Papain, 123 
Papataci fever, 311 
Papilloma, human, 24 

rabbit, see Rabbit papilloma 
Paraminobenzoic acid, 145, 369 
Paracrinkle, potato, 14, 34, 359 

Paracrystals, 94 
Paramecin, 356 

Paramecium aurelia, antigens, 357 
kappa factor, 356, 359 
killer factor, 339 
Parasites, obligate, 3 
Parasitism, 182 
evolution, 347 
genetic, 353 

Particles, counting, 67, 71 
density, 76 
enzymes in, 109 
nonspherical, 75 
size, 56 
spherical, diffusion, 74 

sedimentation, 73 
staining, 58 
visualization, 56 
Passive immunization, 132 
Pasteurella pestis, 366 
Pasteurella tularensis, 366, 372 
Pea mosaic, 14 
Peach mosaic, 14 
Peach rosette, 14 

Peach viruses, temperature effect on, 31 
Peach yellows (little peach), 14, 278 
Penicillin, 145, 179, 223, 369 
Percentage law, 122 
pH effects, 142 
Phagineae, 17 

Phase-contrast microscope, 60 
Phlebotomus papatasii, 311 
Phloem necrosis, 217 
Phony peach, 216 
Photochemical reactions, 148 
Photodynamic action of dyes, 154 
Photoreactivation, bacteriophage, 155, 


Photosynthesis, 217, 219 
Physical agents, effects on viruses, 138 
Phytomonas tumefaciens, 339 
Phytophagineae, 17 
Pigmentation in guinea pig, 359 
Pipettenfehler, 42 
Plant antibodies, 279 
Plant metabolism, 218 
Plant viruses, acquired immunity, 275 
composition, 100 
cross-protection, 275, 278 



Plant viruses, growth, 209/ 
heat treatment, 145 
inclusions, 214, 215 
interference, 275/ 
metabolism, 218 
multiplication in vectors, 317 
necrotic lesions, 21, 22, 46, 47, 304 
seed transmission, 217, 308 
serology, 127 
spread, 216 
symptoms, 26 
tissue cultures, 215 
titration, 46, 54 

transmission, dodder, 34, 308, 309 
graft, 33, 308 
insects, 33, 314 
mechanical, 32, 308 
variation, 296 
Plaque count, 41, 52 

frequency distribution, 49 
proportionality to inoculum, 48 
Plaques, animal viruses, 42 

bacteriophage, 21, 41, 184, 188, 200 
equine encephalomyelitis, 26 
tissue cultures, 25, 42, 226 
Plasmadesmata, 216 
Plasmagenes, 356/ 
Plastids, 214, 272, 354, 360 

mutability, 356 
Pleuropneumonia, 4 
Pneumococcus, type transformation, 351 
Pneumonia virus of mice, 57, 87, 269, 

286, 340 
growth, 222 

hemagglutination, 251, 260 
latency, 37 

Poisson distribution, 49, 165 
Poliomyelitis, 11, 37, 82, 223, 238, 285, 

286, 287, 305 
antibodies, 282 
epidemiology, 310 
Lansing strain, 299 
localization, 25 
passive immunization, 132 
size, 57 
spread, 308 

tissue cultures, 227, 228 
transmission, 309 
vaccination, 133 

Polyhedral diseases, 57, 80, 81, 83 
Polyhedral protein, 110, 111 
Polysaccharide inhibition, 259 

of influenza virus, 146 
Polystyrene latex, 68 
Porthetria dispar, 81 
Potato leaf roll, 5, 14, 218 
Potato paracrinkle, 14, 34, 359 
Potato virus X, 14 

composition, 101 

cross-protection among strains, 276 

morphology, 63, 92 

necrotic response, 24 

size, 57 

synergism, 279 

top necrosis, 217 

transmission, 32 

Potato virus Y, 14, 92, 278, 315 
Potato witches' broom, 36 
Potato yellow dwarf, 14, 57, 92 
Precipitin reaction, 118, 127 
Primary atypical pneumonia, 146 
Prodenia praefica, 80 
Proflavine, 126, 170, 199 
Prophage, see Bacteriophage, prophage 
Protamine, 88 
Proteins, denaturation, 140 

noninfectious, 110 

protective effect, 142 

structure, 98 
Proteus vulgaris, 368 
Protons, 147 
Protozoa, 347, 355 

Provirus, 270, 277, 328, 337, 349, 351 
Pseudorabies (mad itch), 12, 30, 305, 

309, 320 
Psittacosis, 9, 13, 228 

ecology, 302 

inclusions, 232 

transmission, 309 

Psittacosis-lymphogranuloma group, 15, 
18, 245, 298 

chemotherapy, 145, 223 

morphology, 64 

size, 57 

toxicity, 223 
Purification methods, 88 
Purine analogues, 210, 230 



Purity, criteria, 85, 259 
Pyrimidine analogues, 210, 223 

Q fever, 365/ 
Quantum yield, 153 

Rabbit papilloma, 14, 326, 335 
composition, 106 
irradiation, 151, 156, 328 
masking, 327, 361 
serology, 120 
size, 57 

titration, 45, 53 
Rabies, 12, 57, 154, 223, 245, 246, 285, 

286, 299, 305, 308, 309 
fixed strains, 300 
Negri bodies, 25, 230 
street strains, 300 
Radiation, 146 

bacteriophage inactivation, 149, 151, 

153, 156, 165, 180, 194, 195 
effects on influenza virus, 155 
excitation by, 147 
hit theory, 150 
ionizing, 147 

action volume, 150 
direct effect, 150 
indirect effect, 148 
kinetics, 150 
mutagenic effect, 297 
rabbit papilloma, 151, 156, 328 
target theory, 152 
ultraviolet, 152 

quantum yield, 153 
See also Ultraviolet 
visible, 154 

Rail immunization, 280 
Reactivation, 123, 194 

See also Multiplicity reactivation; 

Receptor-destroying enzyme (RDE), 

110, 256/ 

Receptor gradient, 254 
Recombination, see Genetic recombina- 

Reed and Muench method, 43 
Regressive evolution, 4 
Replication, of bacteriophage, 194 
of genetic materials, 158 

Reproduction, 269 

Resistance to viruses, genetic factors, 


Resolving power, 3, 58 
Reverse mutations, 294 
Reversible neutralization, 137 
Riboflavine, 108, 213, 348, 369 
Ribonuclease, 97, 143 
Ribomicleic acid (RNA), 97, 100, 238 
Rice stunt, 14, 317 
Rickettsiae, 3, 347, 364/ 

chemotherapy, 369 

cultivation, 365 

evolution, 370 

in arthropods, 364, 369 

metabolism, 366 

pathogenic classification, 365 

reproduction, 368 

serology, 366 

soluble antigens, 366 

structure, 365 

toxins, 367 

Weil-Felix reaction, 368 
Rickettsia oricntalis, 365/ 
Rickettsia prowazeki, 365/ 
Rift Valley fever, 12, 283, 285 
Rinderpest (cattle plague), 12 
RNA, see Ribomicleic acid 
Rocky Mountain spotted fever, 365/ 
Roentgen unit, 150 
Roller tube, 225 
Root infection, 217, 308 
Rons sarcoma, 13, 31, 107, 246, 324f 

in ducks, 332 

size, 57 

tissue cultures, 227, 235 
Rubella (German measles), 11, 246 

Sacbrood of bees, 13 
Salmonella, bacteriophages, 167 

genetic transfer, 206 

transduction, 352 
Salmonella typhimurium, 206, 208 
Salmonella typhosa, Vt-phages, 167, 206 

Ve-types, 296 

Sandfly fever (papataci fever), 311 
Sarcoma, 322 

See also Rons sarcoma 
Schienenimmunisierung, 280 



Sedimentation, 73/ 
Sedimentation constant, 74 
Sedimentation equilibrium, 76 
Seed transmission, 217, 308 
Separation of virus properties, 120, 135, 


Serologieal cross-reactions, 17, 121, 125 
Serology, 116/ 
rickcttsiae, 366 
See also names of viruses 
Semliki forest virus, 286 
Severe etch, 278 
Shaking, 142 
Sheep dermatitis, 285 
Shigella dysenteriae, 295 
Sigma factor, 357 
Silkworm, 81 

Silkworm jaundice, 13, 320 
induction, 349 
tissue cultures, 237 
Single-burst experiments, 162, 192 
Size of virus particles, 57 
Skin tests, 134 

Smallpox (variola), 5, 24, 29, 230, 298 
vaccination, 132 

Sec also Vaccinia-smallpox group 
Soluble 1 antigen, influenza, 110, 130, 


rickcttsiae, 366 
vaccinia, 110, 130 
Somatic mutations, 338 
Southern bean mosaic, 57, 91, 101 
Specific soluble substances (SSS), 110, 


Spinal fluid, 308 
Spread of viruses, 216, 307 
Squash mosaic virus, 91 
Starch-iodine lesions, 24, 31 
St. Louis encephalitis, 57, 223, 285, 311 
Stokes' law, 73 
Storage, 55 
Strawberry viruses, 87 
Succinic dehyclrogenase, 358 
Sudan III, 331 

Sugar beet curly top, 14, 216, 278, 316 
acquired immunity, 279 
spread, 217 
transmission, 33 
vector, 32, 279, 315 

Sugar beet mosaic, 14 
Sugar cane mosaic, 14 
Sulfhydryl compounds, 144 
Sulfonamides, 145, 223, 369 
Swamp fever (equine infectious ane- 
mia), 12 
Swine influenza, 12, 106, 258, 285, 287 

life cycle, 314 
Swine-pox, 298 
Symbionts, 347, 370 
Synergism, 278, 279, 284 

Tactoids, 94 

Target theory, 152 

Temperature effects, 31, 138 

Teratomas, 334 

Terramycin, 145 

Theiler's virus, see Mouse encephalo- 


Thermal inactivation point, 139 
Thyminc, 97 
Ticks, 311, 370 
Tissue cultures, 7, 215, 224/ 

interference, 283 

plaques, 25, 42, 226 

pure lines, 226 

techniques, 224/ 
Tissue specificity, 7, 227 
Titer, 41 
Titration, 39/ 

bacteriophage, 41, 52 

end point, 43 

errors, 42 

50% infectious dose, 43 

herpes, 249 

incubation period, 45 

precision, 41 

necrotic lesions, 46 

plant viruses, 46, 54 

plaque count, 4 If 

Reed and Muench method, 43 

vaccinia, 40, 51, 53, 249 
Tobacco mosaic, 14, 15, 17, 67, 87, 90 

aucuba strain, 121, 277 

chemical changes, 103, 135, 144 

chemical composition, 101 

crystals, 93 

generation time, 212 

growth, 209/ 



Tobacco mosaic, inactivation, formalin, 

heat, 138 

x-rays, 151 
inclusions, amorphous, 214 

crystalline, iv, 96, 215 
inhibition, 146, 210 
internal structure, 95 
isoelectric point, 88 
layering phenomenon, 95 
metabolic effects, 218 
morphology, 63, 82, 92 
necrotic lesions, 22, 23, 304 
neutralization, 127 
origin, 319 

pepper response to, 22 
ribgrass strain, 277 
root infection, 217, 308 
serological cross-reactions, 121 
size, 57, 93 
synergism, 279 
tissue cultures, 216 
titration, 54 
transmission, 6, 33 
variants, 31, 296 
yellow strains, 27. 276, 278, 297 
X-bodies, 214 
Tobacco necrosis, 14, 17 
composition, 101 
crystals, 90, 92 
growth, 212 

heat inactivation, 138, 141 
noninfectious protein, 110 
root infection, 308 
size, 57 

x-ray effect, 151 

Tobacco ringspot, 14, 23, 138, 276 
Tomato bushy stunt, 14, 15 
composition, 101 
crystals, 86, 90 
growth, 211 
'heat inactivation, 141 
hydration, 91 
morphology, 64 
necrotic lesions, 23 

size, 57, 79 

x-ray effect, 151 
Tomato mosaic, 23 
Tomato spotted wilt, 14, 92 

Top necrosis, 217 

Toxicity, 31, 206, 223, 266, 271 

rickettsiae, 367 
Trachoma, 11, 245 
Transaminase, 366 
Transduction, 206, 352 
Transforming principles, 115, 351 
Transmissible mutagens, 325, 343 
Transmission, artificial, 32/ 

natural, 307/ 

See also names of viruses and virus 


Tricarboxylic acid cycle, 366 
Tristeza of citrus, 14 
Trychocyst, 355 
Trypsin, 143 

Tryptophan as cofactor, 166 
Tulip break, 6, 14, 29 
Tumors, 32 If 

grafting, 322 

fowl, 7, 15, 324/ 

frog kidney, 9, 325, 333 

virus affinity for, 306 

wound, 14, 29, 216, 318, 330 
Turnip yellow mosaic, 14 

composition, 101 

crystals, 91 

noninfectious particles, 110 

size, 57 

structure, 101 

vector, 315 
Typhus, epidemic, 365/ 

murine (endemic), 365/ 
Tyrode's solution, 224 

Ultracentrifugation, 6, 73/ 

Ultracentrifuges, 77 

Ultrafiltrate factor, bacteriophage, 111, 

126, 131, 173 
Ultrafiltration, 6, 70 
Ultraviolet, absorption, 97, 147 

absorption spectra, 98 

origin of life, 346 
Ultraviolet microscope, 58 
Uracil, 97 
Urea, 144 

Vaccination, 6, 132, 133, 248 
Vaccines, 132, 133, 154, 248, 366 



Vaccinia, 6, 11, 331 

antigens, 120, 130 

LS antigens, 110, 130 

chemotherapy, 223 

composition, 106 

elementary bodies, 58, 232 

enzymes, 109 

growth, 247, 270 
inhibition, 230 

hemagglutination, 259 

inclusions (Guarnieri bodies), 233, 
234, 247, 298 

interference, 285, 286 

morphology, 59, 61, 64, 84 

neuro tropic variant, 300 

origin, 348 

serology, 129 

size, 57 

tissue, 226, 228 

titration, 40, 51, 53, 249 

vaccination, 132 

x-ray effects, 151 

Vaccinia-smallpox group, 251, 298, 301 
Variation, 29 If 

animal viruses, 298 

bacteriophagc, 291 

plant viruses, 296 

See also names of viruses and virus 


Varicella, see Chicken pox 
Variegation, 14, 34, 214, 356 
Variola, see Smallpox 
Vectors, 31 Of, 369 
Verniga peruana, 373 
Vi-phages, 167, 206 
Vibrio cholerae, 256 
Viralcs, 17 
Virpid theory, 361 
Virus B, 320 

Virus III, 13, 37, 227, 281, 285, 286, 

Von Magnus phenomenon, 265, 288 

Warts (verrucae), 11, 24, 105, 237, 

238, 325, 331 
Weil-Felix reaction, 368 
West Nile fever, 251, 285, 305 
Wheat mosaic, 14, 308 
Wound tumor, 14, 29, 216, 318, 330 

X-bodies, 214 
X-rays, 147 f, 156 

effects on bacteriophage, 135 

inactivation effects, 151 

Yeast, 357 

Yellow fever, 5, 11, 223 

epidemiology, 313 

immunity, 132 

inclusions, 237 

interference, 283, 285, 286 

jungle, 302, 313 

localization, 29 

multiplication in vectors, 311 

resistance in mice, 30, 305 

size, 57 

strain 17D, 133, 248, 300 

tissue culture, 227 

transmission, 35, 310 

vaccination, 248 
Yellows diseases, 26, 214, 218 

See also names of viruses 
Yolk sac, development, 241 

inoculation, 244 

Zoochlorella, 347 
Zoophagineae, 17